plsql 9i vol2

Oracle9i: Program with PL/SQL

Instructor Guide • Volume 2

40054GC11Production 1.1October 2001D34008

Copyright © Oracle Corporation, 1999, 2000, 2001. All rights reserved.

This documentation contains proprietary information of Oracle Corporation. It isprovided under a license agreement containing restrictions on use and disclosure andis also protected by copyright law. Reverse engineering of the software is prohibited.If this documentation is delivered to a U.S. Government Agency of the Department ofDefense, then it is delivered with Restricted Rights and the following legend isapplicable:

Restricted Rights Legend

Use, duplication or disclosure by the Government is subject to restrictions forcommercial computer software and shall be deemed to be Restricted Rights softwareunder Federal law, as set forth in subparagraph (c)(1)(ii) of DFARS 252.227-7013,Rights in Technical Data and Computer Software (October 1988).

This material or any portion of it may not be copied in any form or by any meanswithout the express prior written permission of Oracle Corporation. Any other copyingis a violation of copyright law and may result in civil and/or criminal penalties.

If this documentation is delivered to a U.S. Government Agency not within theDepartment of Defense, then it is delivered with “Restricted Rights,” as defined inFAR 52.227-14, Rights in Data-General, including Alternate III (June 1987).

The information in this document is subject to change without notice. If you find anyproblems in the documentation, please report them in writing to Education Products,Oracle Corporation, 500 Oracle Parkway, Box SB-6, Redwood Shores, CA 94065.Oracle Corporation does not warrant that this document is error-free.

All references to Oracle and Oracle products are trademarks or registered trademarksof Oracle Corporation.

All other products or company names are used for identification purposes only, andmay be trademarks of their respective owners.

Authors

Nagavalli PataballaPriya Nathan

Technical Contributorsand Reviewers

Anna AtkinsonBryan RobertsCaroline PeredaCesljas ZarcoColey WilliamDaniel GabelDr. Christoph BurandtHakan LindforsHelen RobertsonJohn HoffLachlan WilliamsLaszlo CzinkoczkiLaura PezziniLinda BoldtMarco VerbeekNatarajan SenthilPriya VennapusaRoger AbuzalafRuediger SteffanSarah JonesStefan LindbladSusan Dee

Publisher

Sheryl Domingue

Curriculum Map

I IntroductionCourse Objectives I-2About PL/SQL I-3PL/SQL Environment I-4Benefits of PL/SQL I-5Benefits of Subprograms I-10Invoking Stored Procedures and Functions I-11Summary I-12

1 Declaring VariablesObjectives 1-2PL/SQL Block Structure 1-3Executing Statements and PL/SQL Blocks 1-4Block Types 1-5Program Constructs 1-6Use of Variables 1-7Handling Variables in PL/SQL 1-8Types of Variables 1-9Using iSQL*Plus Variables Within PL/SQL Blocks 1-10Types of Variables 1-11Declaring PL/SQL Variables 1-12Guidelines for Declaring PL/SQL Variables 1-13Naming Rules 1-14Variable Initialization and Keywords 1-15Scalar Data Types 1-17Base Scalar Data Types 1-18Scalar Variable Declarations 1-22The %TYPE Attribute 1-23Declaring Variables with the %TYPE Attribute 1-24Declaring Boolean Variables 1-25Composite Data Types 1-26LOB Data Type Variables 1-27Bind Variables 1-28Using Bind Variables 1-30Referencing Non-PL/SQL Variables 1-31DBMS_OUTPUT.PUT_LINE 1-32Summary 1-33Practice 1 Overview 1-35

Contents

iii

2 Writing Executable StatementsObjectives 2-2PL/SQL Block Syntax and Guidelines 2-3Identifiers 2-5PL/SQL Block Syntax and Guidelines 2-6Commenting Code 2-7SQL Functions in PL/SQL 2-8SQL Functions in PL/SQL: Examples 2-9Data Type Conversion 2-10Nested Blocks and Variable Scope 2-13Identifier Scope 2-15Qualify an Identifier 2-16Determining Variable Scope 2-17Operators in PL/SQL 2-18Programming Guidelines 2-20Indenting Code 2-21Summary 2-22Practice 2 Overview 2-23

3 Interacting with the Oracle ServerObjectives 3-2SQL Statements in PL/SQL 3-3SELECT Statements in PL/SQL 3-4Retrieving Data in PL/SQL 3-7Naming Conventions 3-9Manipulating Data Using PL/SQL 3-10Inserting Data 3-11Updating Data 3-12Deleting Data 3-13Merging Rows 3-14Naming Conventions 3-16SQL Cursor 3-18SQL Cursor Attributes 3-19Transaction Control Statements 3-21Summary 3-22Practice 3 Overview 3-24

iv

4 Writing Control StructuresObjectives 4-2Controlling PL/SQL Flow of Execution 4-3IF Statements 4-4Simple IF Statements 4-5Compound IF Statements 4-6IF-THEN-ELSE Statement Execution Flow 4-7IF-THEN-ELSE Statements 4-8IF-THEN-ELSIF Statement Execution Flow 4-9IF-THEN-ELSIF Statements 4-11CASE Expressions 4-12CASE Expressions: Example 4-13Handling Nulls 4-15Logic Tables 4-16Boolean Conditions 4-17Iterative Control: LOOP Statements 4-18Basic Loops 4-19WHILE Loops 4-21FOR Loops 4-23Guidelines While Using Loops 4-26Nested Loops and Labels 4-27Summary 4-29Practice 4 Overview 4-30

5 Working with Composite Data TypesObjectives 5-2Composite Data Types 5-3PL/SQL Records 5-4Creating a PL/SQL Record 5-5PL/SQL Record Structure 5-7The %ROWTYPE Attribute 5-8Advantages of Using %ROWTYPE 5-10The %ROWTYPE Attribute 5-11INDEX BY Tables 5-13Creating an INDEX BY Table 5-14INDEX BY Table Structure 5-15Creating an INDEX BY Table 5-16Using INDEX BY Table Methods 5-17INDEX BY Table of Records 5-18Example of INDEX BY Table of Records 5-19Summary 5-20Practice 5 Overview 5-21

v

6 Writing Explicit CursorsObjectives 6-2About Cursors 6-3Explicit Cursor Functions 6-4Controlling Explicit Cursors 6-5Declaring the Cursor 6-9Opening the Cursor 6-11Fetching Data from the Cursor 6-12Closing the Cursor 6-14Explicit Cursor Attributes 6-15The %ISOPEN Attribute 6-16Controlling Multiple Fetches 6-17The %NOTFOUND and %ROWCOUNT Attributes 6-18Example 6-20Cursors and Records 6-21Cursor FOR Loops 6-22Cursor FOR Loops Using Subqueries 6-24Summary 6-26Practice 6 Overview 6-27

7 Advanced Explicit Cursor ConceptsObjectives 7-2Cursors with Parameters 7-3The FOR UPDATE Clause 7-5The WHERE CURRENT OF Clause 7-7Cursors with Subqueries 7-9Summary 7-10Practice 7 Overview 7-11

8 Handling ExceptionsObjectives 8-2Handling Exceptions with PL/SQL 8-3Handling Exceptions 8-4Exception Types 8-5Trapping Exceptions 8-6Trapping Exceptions Guidelines 8-7Trapping Predefined Oracle Server Errors 8-8Predefined Exceptions 8-11Trapping Nonpredefined Oracle Server Errors 8-12Nonpredefined Error 8-13Functions for Trapping Exceptions 8-14Trapping User-Defined Exceptions 8-16User-Defined Exceptions 8-17

vi

vii

Calling Environments 8-18Propagating Exceptions 8-19The RAISE_APPLICATION_ERROR Procedure 8-20RAISE_APPLICATION_ERROR 8-22Summary 8-23Practice 8 Overview 8-24

9 Creating ProceduresObjectives 9-2PL/SQL Program Constructs 9-4Overview of Subprograms 9-5Block Structure for Anonymous PL/SQL Blocks 9-6Block Structure for PL/SQL Subprograms 9-7PL/SQL Subprograms 9-8Benefits of Subprograms 9-9Developing Subprograms by Using iSQL*Plus 9-10Invoking Stored Procedures and Functions 9-11What Is a Procedure? 9-12Syntax for Creating Procedures 9-13Developing Procedures 9-14Formal Versus Actual Parameters 9-15Procedural Parameter Modes 9-16Creating Procedures with Parameters 9-17IN Parameters: Example 9-18OUT Parameters: Example 9-19Viewing OUT Parameters 9-21IN OUT Parameters 9-22Viewing IN OUT Parameters 9-23Methods for Passing Parameters 9-24DEFAULT Option for Parameters 9-25Examples of Passing Parameters 9-26Declaring Subprograms 9-27Invoking a Procedure from an Anonymous PL/SQL Block 9-28Invoking a Procedure from Another Procedure 9-29Handled Exceptions 9-30Unhandled Exceptions 9-32Removing Procedures 9-34Summary 9-35Practice 9 Overview 9-37

10 Creating FunctionsObjectives 10-2Overview of Stored Functions 10-3

Syntax for Creating Functions 10-4Creating a Function 10-5Creating a Stored Function by Using iSQL*Plus 10-6Creating a Stored Function by Using iSQL*Plus: Example 10-7Executing Functions 10-8Executing Functions: Example 10-9Advantages of User-Defined Functions in SQL Expressions 10-10Invoking Functions in SQL Expressions: Example 10-11Locations to Call User-Defined Functions 10-12Restrictions on Calling Functions from SQL Expressions 10-13Restrictions on Calling from SQL 10-15Removing Functions 10-16Procedure or Function? 10-17Comparing Procedures and Functions 10-18Benefits of Stored Procedures and Functions 10-19Summary 10-20Practice 10 Overview 10-21

11 Managing SubprogramsObjectives 11-2Required Privileges 11-3Granting Access to Data 11-4Using Invoker's-Rights 11-5Managing Stored PL/SQL Objects 11-6USER_OBJECTS 11-7List All Procedures and Functions 11-8USER_SOURCE Data Dictionary View 11-9List the Code of Procedures and Functions 11-10USER_ERRORS 11-11Detecting Compilation Errors: Example 11-12List Compilation Errors by Using USER_ERRORS 11-13List Compilation Errors by Using SHOW ERRORS 11-14ESCRIBE in iSQL*Plus 11-15Debugging PL/SQL Program Units 11-16Summary 11-17Practice 11 Overview 11-19

12 Creating PackagesObjectives 12-2Overview of Packages 12-3Components of a Package 12-4Referencing Package Objects 12-5Developing a Package 12-6

viii

Creating the Package Specification 12-8Declaring Public Constructs 12-9Creating a Package Specification: Example 12-10Creating the Package Body 12-11Public and Private Constructs 12-12Creating a Package Body: Example 12-13Invoking Package Constructs 12-15Declaring a Bodiless Package 12-17Referencing a Public Variable from a Stand-Alone Procedure 12-18Removing Packages 12-19Guidelines for Developing Packages 12-20Advantages of Packages 12-21Summary 12-23Practice 12 Overview 12-26

13 More Package ConceptsObjectives 13-2Overloading 13-3Overloading: Example 13-5Using Forward Declarations 13-8Creating a One-Time-Only Procedure 13-10Restrictions on Package Functions Used in SQL 13-11User Defined Package: taxes_pack 13-12Invoking a User-Defined Package Function from a SQL Statement 13-13Persistent State of Package Variables: Example 13-14Persistent State of Package Variables 13-15Controlling the Persistent State of a Package Cursor 13-18Executing PACK_CUR 13-20PL/SQL Tables and Records in Packages 13-21Summary 13-22Practice 13 Overview 13-23

14 Oracle Supplied PackagesObjectives 14-2Using Supplied Packages 14-3Using Native Dynamic SQL 14-4Execution Flow 14-5Using the DBMS_SQL Package 14-6Using DBMS_SQL 14-8Using the EXECUTE IMMEDIATE Statement 14-9Dynamic SQL Using EXECUTE IMMEDIATE 14-11Using the DBMS_DDL Package 14-12Using DBMS_JOB for Scheduling 14-13

ix

DBMS_JOB Subprograms 14-14Submitting Jobs 14-15Changing Job Characteristics 14-17Running, Removing, and Breaking Jobs 14-18Viewing Information on Submitted Jobs 14-19Using the DBMS_OUTPUT Package 14-20Interacting with Operating System Files 14-21What Is the UTL_FILE Package? 14-22File Processing Using the UTL_FILE Package 14-23UTL_FILE Procedures and Functions 14-24Exceptions Specific to the UTL_FILE Package 14-25The FOPEN and IS_OPEN Functions 14-26Using UTL_FILE 14-27The UTL_HTTP Package 14-29Using the UTL_HTTP Package 14-30Using the UTL_TCP Package 14-31Oracle-Supplied Packages 14-32Summary 14-37Practice 14 Overview 14-38

15 Manipulating Large ObjectsObjectives 15-2What Is a LOB? 15-3Contrasting LONG and LOB Data Types 15-4Anatomy of a LOB 15-5Internal LOBs 15-6Managing Internal LOBs 15-7What Are BFILEs? 15-8Securing BFILEs 15-9A New Database Object: DIRECTORY 15-10Guidelines for Creating DIRECTORY Objects 15-11Managing BFILEs 15-12Preparing to Use BFILEs 15-13The BFILENAME Function 15-14Loading BFILEs 15-15Migrating from LONG to LOB 15-17The DBMS_LOB Package 15-19DBMS_LOB.READ and DBMS_LOB.WRITE 15-22Adding LOB Columns to a Table 15-23Populating LOB Columns 15-24Updating LOB by Using SQL 15-26Updating LOB by Using DBMS_LOB in PL/SQL 15-27Selecting CLOB Values by Using SQL 15-28

x

Selecting CLOB Values by Using DBMS_LOB 15-29Selecting CLOB Values in PL/SQL 15-30Removing LOBs 15-31Temporary LOBs 15-32Creating a Temporary LOB 15-33Summary 15-34Practice 15 Overview 15-35

16 Creating Database TriggersObjectives 16-2Types of Triggers 16-3Guidelines for Designing Triggers 16-4Database Trigger: Example 16-5Creating DML Triggers 16-6DML Trigger Components 16-7Firing Sequence 16-11Syntax for Creating DML Statement Triggers 16-13Creating DML Statement Triggers 16-14Testing SECURE_EMP 16-15Using Conditional Predicates 16-16Creating a DML Row Trigger 16-17Creating DML Row Triggers 16-18Using OLD and NEW Qualifiers 16-19Using OLD and NEW Qualifiers: Example Using Audit_Emp_Table 16-20Restricting a Row Trigger 16-21INSTEAD OF Triggers 16-22Creating an INSTEAD OF Trigger 16-23Differentiating Between Database Triggers and Stored Procedures 16-28Differentiating Between Database Triggers and Form Builder Triggers 16-29Managing Triggers 16-30DROP TRIGGER Syntax 16-31Trigger Test Cases 16-32Trigger Execution Model and Constraint Checking 16-33Trigger Execution Model and Constraint Checking: Example 16-34A Sample Demonstration for Triggers Using Package Constructs 16-35After Row and After Statement Triggers 16-36Demonstration: VAR_PACK Package Specification 16-37Summary 16-40Practice 16 Overview 16-41

17 More Trigger ConceptsObjectives 17-2Creating Database Triggers 17-3

xi

Creating Triggers on DDL Statements 17-4Creating Triggers on System Events 17-5LOGON and LOGOFF Trigger Example 17-6CALL Statements 17-7Reading Data from a Mutating Table 17-8Mutating Table: Example 17-9Implementing Triggers 17-11Controlling Security Within the Server 17-12Controlling Security with a Database Trigger 17-13Using the Server Facility to Audit Data Operations 17-14Auditing by Using a Trigger 17-15Enforcing Data Integrity Within the Server 17-16Protecting Data Integrity with a Trigger 17-17Enforcing Referential Integrity Within the Server 17-18Protecting Referential Integrity with a Trigger 17-19Replicating a Table Within the Server 17-20Replicating a Table with a Trigger 17-21Computing Derived Data Within the Server 17-22Computing Derived Values with a Trigger 17-23Logging Events with a Trigger 17-24Benefits of Database Triggers 17-26Managing Triggers 17-27Viewing Trigger Information 17-28Using USER_TRIGGERS 17-29Summary 17-31Practice 17 Overview 17-32

18 Managing DependenciesObjectives 18-2Understanding Dependencies 18-3Dependencies 18-4Local Dependencies 18-5A Scenario of Local Dependencies 18-7Displaying Direct Dependencies by Using USER_DEPENDENCIES 18-8Displaying Direct and Indirect Dependencies 18-9Displaying Dependencies 18-10Another Scenario of Local Dependencies 18-11A Scenario of Local Naming Dependencies 18-12Understanding Remote Dependencies 18-13Concepts of Remote Dependencies 18-15REMOTE_DEPENDENCIES_MODE Parameter 18-16Remote Dependencies and Time Stamp Mode 18-17

xii

Remote Procedure B Compiles at 8:00 a.m. 18-19Local Procedure A Compiles at 9:00 a.m. 18-20Execute Procedure A 18-21Remote Procedure B Recompiled at 11:00 a.m. 18-22Signature Mode 18-24Recompiling a PL/SQL Program Unit 18-25Unsuccessful Recompilation 18-26Successful Recompilation 18-27Recompilation of Procedures 18-28Packages and Dependencies 18-29Summary 18-31Practice 18 Overview 18-32

A Practice SolutionsB Table Descriptions and DataC Creating Program Units by Using Procedure BuilderD REF Cursors

IndexAdditional PracticesAdditional Practice SolutionsAdditional Practices: Table Descriptions and Data

xiii

Copyright © Oracle Corporation, 2001. All rights reserved.

Creating Packages

Schedule: Timing Topic

60 minutes Lecture

65 minutes Practice

125 minutes Total

Oracle9i: Program with PL/SQL 12-2

12-2 Copyright © Oracle Corporation, 2001. All rights reserved.

Objectives

After completing this lesson, you should be able todo the following:

• Describe packages and list their possiblecomponents

• Create a package to group together relatedvariables, cursors, constants, exceptions,procedures, and functions

• Designate a package construct as either public orprivate

• Invoke a package construct

• Describe a use for a bodiless package

Lesson Aim

In this lesson you learn what a package is and what its components are. You also learn how to createand use packages.



Overview of Packages

Packages:

• Group logically related PL/SQL types, items, andsubprograms

• Consist of two parts:

– Specification

– Body

• Cannot be invoked, parameterized, or nested

• Allow the Oracle server to read multiple objectsinto memory at once

Packages Overview

Packages bundle related PL/SQL types, items, and subprograms into one container. For example, aHuman Resources package can contain hiring and firing procedures, commission and bonus functions,and tax exemption variables.

A package usually has a specification and a body, stored separately in the database.

The specification is the interface to your applications. It declares the types, variables, constants,exceptions, cursors, and subprograms available for use. The package specification may also includePRAGMAs, which are directives to the compiler.

The body fully defines cursors and subprograms, and so implements the specification.

The package itself cannot be called, parameterized, or nested. Still, the format of a package is similar tothat of a subprogram. Once written and compiled, the contents can be shared by many applications.

When you call a packaged PL/SQL construct for the first time, the whole package is loaded intomemory. Thus, later calls to constructs in the same package require no disk input/output (I/O).


Package Development

You create a package in two parts: first the package specification, and then the package body. Publicpackage constructs are those that are declared in the package specification and defined in the packagebody. Private package constructs are those that are defined solely within the package body.

Note: The Oracle server stores the specification and body of a package separately in the database. Thisenables you to change the definition of a program construct in the package body without causing theOracle server to invalidate other schema objects that call or reference the program construct.


Components of a Package

Procedure Adeclaration

Procedure Adefinition

Procedure Bdefinition

Public variable

Private variable

Public procedure

Private procedure

Public procedure

Local variable

Packagespecification

Packagebody

Scope of the Construct Description Placement within the Package

Public Can be referenced from anyOracle server environment

Declared within the packagespecification and may be definedwithin the package body

Private Can be referenced only byother constructs which arepart of the same package

Declared and defined within thepackage body


Package Development (continued)


Referencing Package Objects


Packagebody




Visibility of the Construct Description

Local A variable defined within a subprogram that is notvisible to external users.Private (local to the package) variable: You candefine variables in a package body. These variablescan be accessed only by other objects in the samepackage. They are not visible to any subprograms orobjects outside of the package.

Global A variable or subprogram that can be referenced(and changed) outside the package and is visible toexternal users. Global package items must bedeclared in the package specification.



Execute

Developing a Package

iSQL*Plus

CodeEditor

Load and run the file.sql2

Source code

P code

Compile

Oracle

1

How to Develop a Package

1. Write the syntax: Enter the code in a text editor and save it as a SQL script file.

2. Compile the code: Run the SQL script file to generate and compile the source code. The sourcecode is compiled into P code.



Developing a Package

• Saving the text of the CREATE PACKAGE statementin two different SQL files facilitates latermodifications to the package.

• A package specification can exist without apackage body, but a package body cannot existwithout a package specification.

How to Develop a Package

There are three basic steps to developing a package, similar to those steps that are used to developa stand-alone procedure.

1. Write the text of the CREATE PACKAGE statement within a SQL script file to create the packagespecification and run the script file. The source code is compiled into P code and is stored withinthe data dictionary.

2. Write the text of the CREATE PACKAGE BODY statement within a SQL script file to create thepackage body and run the script file.The source code is compiled into P code and is also stored within the data dictionary.

3. Invoke any public construct within the package from an Oracle server environment.



CREATE [OR REPLACE] PACKAGE package_nameIS|AS

public type and item declarationssubprogram specifications

END package_name;

Creating the Package Specification

Syntax:

• The REPLACE option drops and recreates thepackage specification.

• Variables declared in the package specification areinitialized to NULL by default.

• All the constructs declared in a packagespecification are visible to users who are grantedprivileges on the package.

How to Create a Package Specification

To create packages, you declare all public constructs within the package specification.

• Specify the REPLACE option when the package specification already exists.

• Initialize a variable with a constant value or formula within the declaration, if required; otherwise,the variable is initialized implicitly to NULL.

Syntax DefinitionP ara m eter D escrip tion

package_name N am e the p ackage

public type anditem declarations

D eclare variables , co ns ta nts, curso rs, exc eptions , or typ es

subprogramspecifications

D eclare th e P L /S Q L sub pro gram s



COMM_PACKAGE package

G_COMM


1

Declaring Public Constructs

RESET_COMMproceduredeclaration

2

Example of a Package SpecificationIn the slide, G_COMM is a public (global) variable, and RESET_COMM is a public procedure.

In the package specification, you declare public variables, public procedures, and public functions.

The public procedures or functions are routines that can be invoked repeatedly by other constructs inthe same package or from outside the package.



Creating a Package Specification:Example

CREATE OR REPLACE PACKAGE comm_package ISg_comm NUMBER := 0.10; --initialized to 0.10PROCEDURE reset_comm(p_comm IN NUMBER);

END comm_package;/

• G_COMM is a global variable and is initialized to 0.10.

• RESET_COMM is a public procedure that isimplemented in the package body.

Package Specification for COMM_PACKAGE

In the slide, the variable G_COMM and the procedure RESET_COMM are public constructs.


Creating the Package Body

To create packages, define all public and private constructs within the package body.

• Specify the REPLACE option when the package body already exists.

• The order in which subprograms are defined within the package body is important: you mustdeclare a variable before another variable or subprogram can refer to it, and you must declare ordefine private subprograms before calling them from other subprograms. It is quite common in thepackage body to see all private variables and subprograms defined first and the publicsubprograms defined last.

Syntax Definition

Define all public and private procedures and functions in the package body.


Creating the Package Body

Syntax:

CREATE [OR REPLACE] PACKAGE BODY package_nameIS|AS

private type and item declarationssubprogram bodies

END package_name;

• The REPLACE option drops and recreates thepackage body.

• Identifiers defined only in the package body areprivate constructs. These are not visible outsidethe package body.

• All private constructs must be declared beforethey are used in the public constructs.

Parameter Description

package_name Is the name of the package

private type anditem declarations

Declares variables, constants, cursors, exceptions, or types

subprogram bodies Defines the PL/SQL subprograms, public and private



RESET_COMMprocedure declaration

VALIDATE_COMMfunction definition


Packagebody

1

3

2RESET_COMMprocedure definition

COMM_PACKAGE package

Public and Private Constructs

G_COMM

2

Create a Package Body Example

In the slide on this page:

• 1 is a public (global) variable

• 2 is a public procedure

• 3 is a private function

You can define a private procedure or function to modularize and clarify the code of public proceduresand functions.

Note: In the slide, the private function is shown above the public procedure. When you are coding thepackage body, the definition of the private function has to be above the definition of the publicprocedure.

Only subprograms and cursors declarations without body in a package specification have an underlyingimplementation in the package body. So if a specification declares only types, constants, variables,exceptions, and call specifications, the package body is unnecessary. However, the body can still be usedto initialize items declared in the package specification.



Creating a Package Body: Example

CREATE OR REPLACE PACKAGE BODY comm_packageIS

FUNCTION validate_comm (p_comm IN NUMBER)RETURN BOOLEAN

ISv_max_comm NUMBER;

BEGINSELECT MAX(commission_pct)INTO v_max_commFROM employees;

IF p_comm > v_max_comm THEN RETURN(FALSE);ELSE RETURN(TRUE);END IF;

END validate_comm;...

comm_pack.sql

Package Body for COMM_PACKAGE

Define a function to validate the commission. The commission may not be greater than the highestcommission among all existing employees.

Instructor NoteThe code in the slide is continued on the following page.

A cursor can be declared and defined inside the package specification without requiring an underlyingimplementation in the package body. Cursors that are specified only with a return type should beimplemented inside the package body. Example:CREATE OR REPLACE PACKAGE TESTCURSOR ISCURSOR C RETURN employees%ROWTYPE;

...END;CREATE OR REPLACE PACKAGE BODY TESTCURSOR IS

CURSOR C RETURN employees%ROWTYPE IS SELECT * FROM employees;...END;



PROCEDURE reset_comm (p_comm IN NUMBER)ISBEGINIF validate_comm(p_comm)THEN g_comm:=p_comm; --reset global variable

ELSERAISE_APPLICATION_ERROR(-20210,'Invalid commission');

END IF;END reset_comm;

END comm_package;/

Creating a Package Body: Example

comm_pack.sql

Package Body for COMM_PACKAGE (continued)

Define a procedure that enables you to reset and validate the prevailing commission.



Invoking Package Constructs

Example 1: Invoke a function from a procedure withinthe same package.

CREATE OR REPLACE PACKAGE BODY comm_package IS. . .

PROCEDURE reset_comm(p_comm IN NUMBER)

ISBEGINIF validate_comm(p_comm)

THEN g_comm := p_comm;ELSE

RAISE_APPLICATION_ERROR(-20210, 'Invalid commission');

END IF;END reset_comm;

END comm_package;


After the package is stored in the database, you can invoke a package construct within the package orfrom outside the package, depending on whether the construct is private or public.

When you invoke a package procedure or function from within the same package, you do not need toqualify its name.

Example 1

Call the VALIDATE_COMM function from the RESET_COMM procedure. Both subprograms are in theCOMM_PACKAGE package.


Invoking Package Constructs (continued)

When you invoke a package procedure or function from outside the package, you must qualify itsname with the name of the package.

Example 2

Call the RESET_COMM procedure from iSQL*Plus, making the prevailing commission 0.15 for theuser session.

Example 3

Call the RESET_COMM procedure that is located in the SCOTT schema from iSQL*Plus, making theprevailing commission 0.15 for the user session.

Example 4

Call the RESET_COMM procedure that is located in a remote database that is determined by thedatabase link named NY from iSQL*Plus, making the prevailing commission 0.15 for the user session.

Adhere to normal naming conventions for invoking a procedure in a different schema, or in a differentdatabase on another node.

Instructor Note

In example 4, the name of the schema from which the package is accessed is included in the databaselink.



Example 2: Invoke a package procedure from iSQL*Plus.

Example 3: Invoke a package procedure in a differentschema.

Example 4: Invoke a package procedure in a remotedatabase.

EXECUTE comm_package.reset_comm(0.15)

EXECUTE scott.comm_package.reset_comm(0.15)

EXECUTE comm_package.reset_comm@ny(0.15)



Declaring a Bodiless Package

CREATE OR REPLACE PACKAGE global_consts ISmile_2_kilo CONSTANT NUMBER := 1.6093;kilo_2_mile CONSTANT NUMBER := 0.6214;yard_2_meter CONSTANT NUMBER := 0.9144;meter_2_yard CONSTANT NUMBER := 1.0936;

END global_consts;/

EXECUTE DBMS_OUTPUT.PUT_LINE('20 miles = '||20*global_consts.mile_2_kilo||' km')

Declaring a Bodiless Package

You can declare public (global) variables that exist for the duration of the user session. You cancreate a package specification that does not need a package body. As discussed earlier in thislesson, if a specification declares only types, constants, variables, exceptions, and callspecifications, the package body is unnecessary.

Example

In the example in the slide, a package specification containing several conversion rates is defined.All the global identifiers are declared as constants.

A package body is not required to support this package specification because implementationdetails are not required for any of the constructs of the package specification.

Instructor Note

This slide highlights two concepts:

• A package specification without a package body

• The uses of global variables



Referencing a Public Variable froma Stand-Alone Procedure

Example:CREATE OR REPLACE PROCEDURE meter_to_yard

(p_meter IN NUMBER, p_yard OUT NUMBER)ISBEGIN

p_yard := p_meter * global_consts.meter_2_yard;END meter_to_yard;/VARIABLE yard NUMBEREXECUTE meter_to_yard (1, :yard)PRINT yard

ExampleUse the METER_TO_YARD procedure to convert meters to yards, using the conversion rate packaged inGLOBAL_CONSTS.

When you reference a variable, cursor, constant, or exception from outside the package, you mustqualify its name with the name of the package.



To remove the package specification and the body,use the following syntax:

To remove the package body, use the following syntax:

DROP PACKAGE package_name;

Removing Packages

DROP PACKAGE BODY package_name;

Removing a Package

When a package is no longer required, you can use a SQL statement in iSQL*Plus to drop it. Apackage has two parts, so you can drop the whole package or just the package body and retain thepackage specification.



Guidelines for Developing Packages

• Construct packages for general use.

• Define the package specification before the body.

• The package specification should contain onlythose constructs that you want to be public.

• Place items in the declaration part of the packagebody when you must maintain them throughouta session or across transactions.

• Changes to the package specification requirerecompilation of each referencing subprogram.

• The package specification should contain as fewconstructs as possible.

Guidelines for Writing Packages

Keep your packages as general as possible so that they can be reused in future applications. Also, avoidwriting packages that duplicate features provided by the Oracle server.

Package specifications reflect the design of your application, so define them before defining the packagebodies.

The package specification should contain only those constructs that must be visible to users of thepackage. That way other developers cannot misuse the package by basing code on irrelevant details.

Place items in the declaration part of the package body when you must maintain them throughout asession or across transactions. For example, declare a variable called NUMBER_EMPLOYED as a privatevariable, if each call to a procedure that uses the variable needs to be maintained. When declared as aglobal variable in the package specification, the value of that global variable gets initialized in a sessionthe first time a construct from the package is invoked.

Changes to the package body do not require recompilation of dependent constructs, whereas changes tothe package specification require recompilation of every stored subprogram that references the package.To reduce the need for recompiling when code is changed, place as few constructs as possible in apackage specification.



Advantages of Packages

• Modularity: Encapsulate related constructs.

• Easier application design: Code and compilespecification and body separately.

• Hiding information:

– Only the declarations in the packagespecification are visible and accessible toapplications.

– Private constructs in the package body arehidden and inaccessible.

– All coding is hidden in the package body.

Advantages of Using Packages

Packages provide an alternative to creating procedures and functions as stand-alone schema objects, andthey offer several benefits.

Modularity

You encapsulate logically related programming structures in a named module. Each package is easy tounderstand, and the interface between packages is simple, clear, and well defined.

Easier Application Design

All you need initially is the interface information in the package specification. You can code andcompile a specification without its body. Then stored subprograms that reference the package cancompile as well. You need not define the package body fully until you are ready to complete theapplication.

Hiding Information

You can decide which constructs are public (visible and accessible) or private (hidden and inaccessible).Only the declarations in the package specification are visible and accessible to applications. The packagebody hides the definition of the private constructs so that only the package is affected (not yourapplication or any calling programs) if the definition changes. This enables you to change theimplementation without having to recompile calling programs. Also, by hiding implementation detailsfrom users, you protect the integrity of the package.



Advantages of Packages

• Added functionality: Persistency of variablesand cursors

• Better performance:

– The entire package is loaded into memorywhen the package is first referenced.

– There is only one copy in memory for all users.

– The dependency hierarchy is simplified.

• Overloading: Multiple subprograms of thesame name

Advantages of Using Packages (continued)

Added Functionality

Packaged public variables and cursors persist for the duration of a session. Thus, they can be shared byall subprograms that execute in the environment. They also enable you to maintain data acrosstransactions without having to store it in the database. Private constructs also persist for the duration ofthe session, but can only be accessed within the package.

Better Performance

When you call a packaged subprogram the first time, the entire package is loaded into memory. Thisway, later calls to related subprograms in the package require no further disk I/O. Packagedsubprograms also stop cascading dependencies and so avoid unnecessary compilation.

Overloading

With packages you can overload procedures and functions, which means you can create multiplesubprograms with the same name in the same package, each taking parameters of different number ordatatype.

Instructor Note

Dependencies are covered in a later lesson.



Summary

In this lesson, you should have learned how to:

• Improve organization, management, security, andperformance by using packages

• Group related procedures and functions togetherin a package

• Change a package body without affecting apackage specification

• Grant security access to the entire package

Summary

You group related procedures and function together into a package. Packages improve organization,management, security, and performance.

A package consists of package specification and a package body. You can change a package bodywithout affecting its package specification.



Summary


• Hide the source code from users

• Load the entire package into memory on thefirst call

• Reduce disk access for subsequent calls

• Provide identifiers for the user session

Summary (continued)

Packages enable you to hide source code from users. When you invoke a package for the first time, theentire package is loaded into memory. This reduces the disk access for subsequent calls.



Command

CREATE [OR REPLACE] PACKAGE

CREATE [OR REPLACE] PACKAGEBODY

DROP PACKAGE

DROP PACKAGE BODY

Task

Create (or modify) an existingpackage specification

Create (or modify) an existingpackage body

Remove both the packagespecification and the package body

Remove the package body only

Summary

Summary (continued)

You can create, delete, and modify packages. You can remove both package specification and bodyby using the DROP PACKAGE command. You can drop the package body without affecting itsspecification.



Practice 12 Overview

This practice covers the following topics:

• Creating packages

• Invoking package program units


In this practice, you will create package specifications and package bodies. You will invoke theconstructs in the packages, using sample data.


Practice 121. Create a package specification and body called JOB_PACK. (You can save the package body

and specification in two separate files.) This package contains your ADD_JOB, UPD_JOB, andDEL_JOB procedures, as well as your Q_JOB function.

Note: Use the code in your previously saved script files when creating the package.

a. Make all the constructs public.

Note: Consider whether you still need the stand-alone procedures and functions you justpackaged.

b. Invoke your ADD_JOB procedure by passing values IT_SYSAN and SYSTEMSANALYST as parameters.

c. Query the JOBS table to see the result.

2. Create and invoke a package that contains private and public constructs.

a. Create a package specification and package body called EMP_PACK that contains yourNEW_EMP procedure as a public construct, and your VALID_DEPTID function as aprivate construct. (You can save the specification and body into separate files.)

b. Invoke the NEW_EMP procedure, using 15 as a department number. Because thedepartment ID 15 does not exist in the DEPARTMENTS table, you should get an errormessage as specified in the exception handler of your procedure.

c. Invoke the NEW_EMP procedure, using an existing department ID 80.

If you have time:

3. a. Create a package called CHK_PACK that contains the procedures CHK_HIREDATE andCHK_DEPT_MGR. Make both constructs public. (You can save the specification and bodyinto separate files.) The procedure CHK_HIREDATE checks whether an employee’s hiredate is within the following range: [SYSDATE - 50 years, SYSDATE + 3 months].

Note:

• If the date is invalid, you should raise an application error with an appropriatemessage indicating why the date value is not acceptable.

• Make sure the time component in the date value is ignored.

• Use a constant to refer to the 50 years boundary.

• A null value for the hire date should be treated as an invalid hire date.The procedure CHK_DEPT_MGR checks the department and manager combination for agiven employee. The CHK_DEPT_MGR procedure accepts an employee ID and a managerID. The procedure checks that the manager and employee work in the same department.The procedure also checks that the job title of the manager ID provided isMANAGER.

Note: If the department ID and manager combination is invalid, you should raise anapplication error with an appropriate message.


Practice 12 (continued)b. Test the CHK_HIREDATE procedure with the following command:

EXECUTE chk_pack.chk_hiredate('01-JAN-47')

What happens, and why?

c. Test the CHK_HIREDATE procedure with the following command:

EXECUTE chk_pack.chk_hiredate(NULL)


d. Test the CHK_DEPT_MGR procedure with the following command:

EXECUTE chk_pack.chk_dept_mgr(117,100)


Schedule: Timing Topic

50 minutes Lecture

45 minutes Practice

95 minutes Total


More Package Concepts



Objectives


• Write packages that use the overloading feature

• Describe errors with mutually referentialsubprograms

• Initialize variables with a one-time-only procedure

• Identify persistent states

Lesson Aim

This lesson introduces more advanced features of PL/SQL, including overloading, forward referencing, aone-time-only procedure, and the persistency of variables, constants, exceptions, and cursors. It also looksat the effect of packaging functions that are used in SQL statements.



Overloading

• Enables you to use the same name for differentsubprograms inside a PL/SQL block, asubprogram, or a package

• Requires the formal parameters of thesubprograms to differ in number, order, or datatype family

• Enables you to build more flexibility because auser or application is not restricted by the specificdata type or number of formal parameters

Note: Only local or packaged subprograms can beoverloaded. You cannot overload stand-alonesubprograms.

OverloadingThis feature enables you to define different subprograms with the same name. You can distinguish thesubprograms both by name and by parameters. Sometimes the processing in two subprograms is thesame, but the parameters passed to them varies. In that case it is logical to give them the same name.PL/SQL determines which subprogram is called by checking its formal parameters. Only local orpackaged subprograms can be overloaded. Stand-alone subprograms cannot be overloaded.RestrictionsYou cannot overload:

• Two subprograms if their formal parameters differ only in data type and the different data typesare in the same family (NUMBER and DECIMAL belong to the same family)

• Two subprograms if their formal parameters differ only in subtype and the different subtypes arebased on types in the same family (VARCHAR and STRING are PL/SQL subtypes ofVARCHAR2)

• Two functions that differ only in return type, even if the types are in different familiesYou get a run-time error when you overload subprograms with the above features.Note: The above restrictions apply if the names of the parameters are also the same. If you usedifferent names for the parameters, then you can invoke the subprograms by using named notation forthe parameters.


Overloading (continued)

Resolving CallsThe compiler tries to find a declaration that matches the call. It searches first in the current scope andthen, if necessary, in successive enclosing scopes. The compiler stops searching if it finds one or moresubprogram declarations in which the name matches the name of the called subprogram. For like-namedsubprograms at the same level of scope, the compiler needs an exact match in number, order, and datatype between the actual and formal parameters.

Instructor Note

You can overload two subprograms if their formal parameters differ only in name or parameter mode.You can also overload two subprograms with the restrictions specified in the first bullet point under"restrictions" on page 13-3. If you invoke the overloaded method with named parameter notation, youdo not get a run-time error.

For example, consider the package OVERLOAD, which contains two overloaded procedures. Theprocedures differ only in the name of the parameters:CREATE OR REPLACE PACKAGE overload ISprocedure p (x number);procedure p ( n number);

END;/After you create the package body, execute procedure P as follows:

EXECUTE overload.p(4)overload.p(4); END;*

ERROR at line 1:ORA-06550: line 1, column 7:PLS-00307: too many declarations of 'P' match this callORA-06550: line 1, column 7:PL/SQL: Statement ignored

Now, invoke the procedure with named notation:

EXECUTE overload.p(x => 4)

EXECUTE overload.p(n=> 3)

PL/SQL procedure successfully completed.



Overloading: Example

CREATE OR REPLACE PACKAGE over_pack

IS

PROCEDURE add_dept

(p_deptno IN departments.department_id%TYPE,

p_name IN departments.department_name%TYPEDEFAULT 'unknown',

p_loc IN departments.location_id%TYPE DEFAULT 0);

PROCEDURE add_dept

(p_name IN departments.department_name%TYPEDEFAULT 'unknown',

p_loc IN departments.location_id%TYPE DEFAULT 0);

END over_pack;

/

over_pack.sql


The slide shows the package specification of a package with overloaded procedures.

The package contains ADD_DEPT as the name of two overloaded procedures. The first definition takesthree parameters to be able to insert a new department to the department table. The second definitiontakes only two parameters, because the department ID is populated through a sequence.

Instructor Note (for page 13-6)

Defaults for parameters are the most common cause of overload errors. Ask the class what happens ifthe parameter p_deptno has a default value in the first declaration of the procedure ADD_DEPT andthe procedure ADD_DEPT is invoked with no arguments. You receive an error

PLS-00307: too many declarations of 'ADD_DEPT' match this call.




CREATE OR REPLACE PACKAGE BODY over_pack ISPROCEDURE add_dept(p_deptno IN departments.department_id%TYPE,p_name IN departments.department_name%TYPE DEFAULT 'unknown',p_loc IN departments.location_id%TYPE DEFAULT 0)

ISBEGININSERT INTO departments (department_id,

department_name, location_id)VALUES (p_deptno, p_name, p_loc);

END add_dept;PROCEDURE add_dept(p_name IN departments.department_name%TYPE DEFAULT 'unknown',p_loc IN departments.location_id%TYPE DEFAULT 0)

ISBEGININSERT INTO departments (department_id,

department_name, location_id)VALUES (departments_seq.NEXTVAL, p_name, p_loc);

END add_dept;END over_pack;

/

over_pack_body.sql

Overloading Example (continued)If you call ADD_DEPT with an explicitly provided department ID, PL/SQL uses the first version of theprocedure. If you call ADD_DEPT with no department ID, PL/SQL uses the second version.

EXECUTE over_pack.add_dept (980,'Education',2500)

EXECUTE over_pack.add_dept ('Training', 2400)

SELECT * FROM departments

WHERE department_id = 980;

SELECT * FROM departments

WHERE department_name = 'Training';




• Most built-in functions are overloaded.• For example, see the TO_CHAR function of the

STANDARD package.

• If you redeclare a built-in subprogram in a PL/SQLprogram, your local declaration overrides theglobal declaration.

FUNCTION TO_CHAR (p1 DATE) RETURN VARCHAR2;FUNCTION TO_CHAR (p2 NUMBER) RETURN VARCHAR2;FUNCTION TO_CHAR (p1 DATE, P2 VARCHAR2) RETURN VARCHAR2;FUNCTION TO_CHAR (p1 NUMBER, P2 VARCHAR2) RETURN VARCHAR2;

Overloading Example (continued)Most built-in functions are overloaded. For example, the function TO_CHAR in the package STANDARDhas four different declarations, as shown in the slide. The function can take either the DATE or theNUMBER data type and convert it to the character data type. The format into which the date or numberhas to be converted can also be specified in the function call.

If you redeclare a built-in subprogram in another PL/SQL program, your local declaration overrides thestandard or built-in subprogram. To be able to access the built-in subprogram, you need to qualify it withits package name. For example, if you redeclare the TO_CHAR function, to access the built-in functionyou refer it as: STANDARD.TO_CHAR.

If you redeclare a built-in subprogram as a stand-alone subprogram, to be able to access yoursubprogram you need to qualify it with your schema name, for example, SCOTT.TO_CHAR.

Instructor NoteThere is a standard.sql script file in the RDBMS\ADMIN directory in the Oracle home directory.Using this script file, you can demonstrate overloading subprograms such as TO_CHAR or NVL from theSTANDARD package.



Using Forward Declarations

You must declare identifiers before referencing them.

CREATE OR REPLACE PACKAGE BODY forward_packIS

PROCEDURE award_bonus(. . .)ISBEGINcalc_rating(. . .); --illegal reference

END;

PROCEDURE calc_rating(. . .)ISBEGIN...

END;

END forward_pack;/


PL/SQL does not allow forward references. You must declare an identifier before using it. Therefore, asubprogram must be declared before calling it.

In the example in the slide, the procedure CALC_RATING cannot be referenced because it has not yetbeen declared. You can solve the illegal reference problem by reversing the order of the two procedures.However, this easy solution does not always work. Suppose the procedures call each other or youabsolutely want to define them in alphabetical order.

PL/SQL enables for a special subprogram declaration called a forward declaration. It consists of thesubprogram specification terminated by a semicolon. You can use forward declarations to do thefollowing:

• Define subprograms in logical or alphabetical order

• Define mutually recursive subprograms

• Group subprograms in a package

Mutually recursive programs are programs that call each other directly or indirectly.

Note: If you receive a compilation error that CALC_RATING is undefined, it is only a problem ifCALC_RATING is a private packaged procedure. If CALC_RATING is declared in the packagespecification, the reference to the public procedure is resolved by the compiler.




CREATE OR REPLACE PACKAGE BODY forward_packIS

PROCEDURE calc_rating(. . .); -- forward declaration

PROCEDURE award_bonus(. . .)IS -- subprograms definedBEGIN -- in alphabetical ordercalc_rating(. . .);

. . .END;

PROCEDURE calc_rating(. . .)ISBEGIN. . .

END;

END forward_pack;/

Using Forward Declarations (continued)

• The formal parameter list must appear in both the forward declaration and the subprogram body.

• The subprogram body can appear anywhere after the forward declaration, but both must appear inthe same program unit.

Forward Declarations and Packages

Forward declarations typically let you group related subprograms in a package. The subprogramspecifications go in the package specification, and the subprogram bodies go in the package body,where they are invisible to the applications. In this way, packages enable you to hide implementationdetails.

Instructor Note

A forward declaration has the same syntax as is used in the package specification for a publicprocedure or function.



Creating a One-Time-Only Procedure

CREATE OR REPLACE PACKAGE taxesIS

tax NUMBER;... -- declare all public procedures/functions

END taxes;/

CREATE OR REPLACE PACKAGE BODY taxesIS

... -- declare all private variables

... -- define public/private procedures/functionsBEGIN

SELECT rate_valueINTO taxFROM tax_ratesWHERE rate_name = 'TAX';

END taxes;/

Define an Automatic, One-Time-Only Procedure

A one-time-only procedure is executed only once, when the package is first invoked within the usersession. In the preceding slide, the current value for TAX is set to the value in the TAX_RATES tablethe first time the TAXES package is referenced.

Note: Initialize public or private variables with an automatic, one-time-only procedure when thederivation is too complex to embed within the variable declaration. In this case, do not initialize thevariable in the declaration, because the value is reset by the one-time-only procedure.

The keyword END is not used at the end of a one-time-only procedure. Observe that in the example inthe slide, there is no END at the end of the one-time-only procedure.

Instructor NoteIn the example, TAX could be a global or local (to the package) variable; the one-time-only procedurewould be the same.



Restrictions on Package FunctionsUsed in SQL

A function called from:

• A query or DML statement can not end the currenttransaction, create or roll back to a savepoint, orALTER the system or session.

• A query statement or a parallelized DML statementcan not execute a DML statement or modify thedatabase.

• A DML statement can not read or modify theparticular table being modified by that DMLstatement.

Note: Calls to subprograms that break the aboverestrictions are not allowed.

Controlling Side EffectsFor the Oracle server to execute a SQL statement that calls a stored function, it must know the purity levelof a stored functions, that is, whether the functions are free of side effects. Side effects are changes todatabase tables or public packaged variables (those declared in a package specification). Side effects coulddelay the execution of a query, yield order-dependent (therefore indeterminate) results, or require that thepackage state variables be maintained across user sessions. Various side effects are not allowed when afunction is called from a SQL query or DML statement. Therefore, the following restrictions apply tostored functions called from SQL expressions:

• A function called from a query or DML statement can not end the current transaction, create or rollback to a savepoint, or alter the system or session

• A function called from a query statement or from a parallelized DML statement can not execute aDML statement or otherwise modify the database

• A function called from a DML statement can not read or modify the particular table being modifiedby that DML statement

Note: In releases prior to Oracle8i, the purity checking used to be performed during compilation time, byincluding the PRAGMA RESTRICT_REFERENCES compiler directive in the package specification. Butfrom Oracle8i, a user-written function can be called from a SQL statement without any compile-timechecking of its purity. You can use PRAGMA RESTRICT_REFERENCES to ask the PL/SQL compiler toverify that a function has only the side effects that you expect. SQL statements, package variable accesses,or calls to functions that violate the declared restrictions continue to raise PL/SQL compilation errors tohelp you isolate the code that has unintended effects.Note: The restrictions on functions discussed above are the same as those discussed in the lesson “CreatingFunctions.”



CREATE OR REPLACE PACKAGE taxes_packIS

FUNCTION tax (p_value IN NUMBER) RETURN NUMBER;END taxes_pack;/

User Defined Package: taxes_pack

CREATE OR REPLACE PACKAGE BODY taxes_packIS

FUNCTION tax (p_value IN NUMBER) RETURN NUMBERISv_rate NUMBER := 0.08;

BEGINRETURN (p_value * v_rate);

END tax;END taxes_pack;/

ExampleEncapsulate the function TAX in the package TAXES_PACK. The function is called from SQL statementson remote databases.

Instructor Note

The example in the slide is a very simple demonstration, as it contains only one function in the package,which is not a common occurrence. This shows an example of a package function that can be used in aSQL statement.

In versions prior to Oracle8i, if your package contains a one-time-only procedure, you must also assertthe purity level of that section. The compiler needs to know whether your one-time-only proceduremodifies package variables or database information. You do this in your package specification byproviding the PRAGMA RESTRICT_REFERENCES clause with the name of the package and allapplicable state parameters. You also use PRAGMA RESTRICT_REFERENCES for calling externalprocedures. Calling external procedures is discussed in the course Advanced PL/SQL.



Invoking a User-Defined Package Functionfrom a SQL Statement

SELECT taxes_pack.tax(salary), salary, last_nameFROM employees;

…

Calling Package Functions

You call PL/SQL functions the same way that you call built-in SQL functions.

Example

Call the TAX function (in the TAXES_PACK package) from a SELECT statement.

Note: If you are using Oracle versions prior to 8i, you need to assert the purity level of the function inthe package specification by using PRAGMA RESTRICT_REFERENCES. If this is not specified, youget an error message saying that the function TAX does not guarantee that it will not update thedatabase while invoking the package function in a query.



Persistent State of PackageVariables: Example

CREATE OR REPLACE PACKAGE comm_package ISg_comm NUMBER := 10; --initialized to 10PROCEDURE reset_comm (p_comm IN NUMBER);

END comm_package;/

CREATE OR REPLACE PACKAGE BODY comm_package ISFUNCTION validate_comm (p_comm IN NUMBER)

RETURN BOOLEANIS v_max_comm NUMBER;BEGIN... -- validates commission to be less than maximum

-- commission in the tableEND validate_comm;PROCEDURE reset_comm (p_comm IN NUMBER)IS BEGIN

... -- calls validate_comm with specified valueEND reset_comm;

END comm_package;/

Persistent State of Package VariablesThis sample package illustrates the persistent state of package variables. The VALIDATE_COMMfunction validates commission to be no more than maximum currently earned. The RESET_COMMprocedure invokes the VALIDATE_COMM function. If you try to reset the commission to be higherthan the maximum, the exception RAISE_APPLICATION_ERROR is raised. On the next page, theRESET_COMM procedure is used in the example.

Note: Refer to page 13 of lesson 12 for the code of the VALIDATE_COMM function and theRESET_COMM procedure. In the VALIDATE_COMM function, the maximum salary from theEMPLOYEES table is selected into the variable V_MAXSAL. Once the variable is assigned a value, thevalue persists in the session until it is modified again. The example in the following slide shows howthe value of a global package variable persists for a session.


Illustrate how the variable are stored in the PGA for the two different sessions.



Persistent State of Package VariablesTime9:00 EXECUTE

comm_package.reset_comm(0.25)max_comm=0.4 > 0.25g_comm = 0.259:30

INSERT INTO employees(last_name, commission_pct)VALUES ('Madonna', 0.8);max_comm=0.8

9:35EXECUTEcomm_package.reset_comm(0.5)

max_comm=0.8 > 0.5g_comm = 0.5

Scott Jones

Controlling the Persistent State of a Package VariableYou can keep track of the state of a package variable or cursor, which persists throughout the usersession, from the time the user first references the variable or cursor to the time the user disconnects.

1. Initialize the variable within its declaration or within an automatic, one-time-only procedure.

2. Change the value of the variable by means of package procedures.

3. The value of the variable is released when the user disconnects.

The sequence of steps in the slide shows how the state of a package variable persists.9:00: When Scott invoked the procedure RESET_COMM with a commission percentage value 0.25, theglobal variable G_COMM was initialized to 10 in his session. The value 0.25 was validated with themaximum commission percentage value 0.4 (obtained from the EMPLOYEES table). Because 0.25 isless than 0.4, the global variable was set to 0.25.9:30: Jones inserted a new row into EMPLOYEES table with commission percentage value 0.8.

9:35: Jones invoked the procedure RESET_COMM with a commission percentage value 0.5. The globalvariable G_COMM was initialized to 10 in his session. The value 0.5 was validated with the maximumcommission percentage value 0.8 (because the new row has 0.8). Because 0.5 is less than 0.8, theglobal variable was set to 0.5.







max_comm=0.8 > 0.5g_comm = 0.510:00 EXECUTE

comm_package.reset_comm(0.6)max_comm=0.4 < 0.6 INVALID

11:00 ROLLBACK;

11:01 EXIT

Scott Jones

Controlling the Persistent State of a Package Variable (continued)10:00: Scott invoked the procedure with commission percentage value of 0.6. This value is more thanthe maximum commission percentage 0.4 (Scott could not see new value because Jones did notcomplete the transaction). Hence, it was invalid.







max_comm=0.8 > 0.5g_comm = 0.510:00 EXECUTE

comm_package.reset_comm(0.6)max_comm=0.4 < 0.6 INVALID

11:00 ROLLBACK;

11:01 EXIT

11:45 Logged In again. g_comm = 10,max_comm=0.4

12:00 EXECUTEcomm_package.reset_comm(0.25)

Scott Jones

VALID

Controlling the Persistent State of a Package Variable (continued)11:00 to 12:00: Jones rolled back the transaction and exited the session. The global value wasinitialized to 10 when he logged in at 11:45. The procedure was successful because the new value0.25 is less than the maximum value 0.4.



Controlling the Persistent State of aPackage Cursor

CREATE OR REPLACE PACKAGE pack_cur

IS

CURSOR c1 IS SELECT employee_id

FROM employees

ORDER BY employee_id DESC;

PROCEDURE proc1_3rows;

PROCEDURE proc4_6rows;

END pack_cur;

/

Example:

Controlling the Persistent State of a Package Cursor

Example

Use the following steps to control a public cursor:

1. Declare the public (global) cursor in the package specification.

2. Open the cursor and fetch successive rows from the cursor, using one (public) packagedprocedure, PROC1_3ROWS.

3. Continue to fetch successive rows from the cursor, and then close the cursor by using another(public) packaged procedure, PROC4_6ROWS.

The slide shows the package specification for PACK_CUR.



CREATE OR REPLACE PACKAGE BODY pack_cur ISv_empno NUMBER;PROCEDURE proc1_3rows ISBEGINOPEN c1;LOOPFETCH c1 INTO v_empno;DBMS_OUTPUT.PUT_LINE('Id :' ||(v_empno));EXIT WHEN c1%ROWCOUNT >= 3;END LOOP;

END proc1_3rows;PROCEDURE proc4_6rows ISBEGINLOOPFETCH c1 INTO v_empno;DBMS_OUTPUT.PUT_LINE('Id :' ||(v_empno));EXIT WHEN c1%ROWCOUNT >= 6;END LOOP;CLOSE c1;

END proc4_6rows;END pack_cur;/

Controlling the Persistent State of aPackage Cursor

Controlling the Persistent State of a Package Cursor (continued)

Example

The slide on this page shows the package body for PACK_CUR to support the package specification.In the package body:

1. Open the cursor and fetch successive rows from the cursor by using one packaged procedure,PROC1_3ROWS.

2. Continue to fetch successive rows from the cursor and close the cursor, using anotherpackaged procedure, PROC4_6ROWS.

Instructor NoteIt is not a good practice to leave cursors open when you exit the the PROC1_3ROWS procedure.

If the procedure PROC4_6ROWS is invoked prior to the PROC1_3ROWS procedure, you get anexception because the cursor is not opened in the PROC4_6ROWS procedure.



Executing PACK_CUR

SET SERVEROUTPUT ON

EXECUTE pack_cur.proc1_3rows

EXECUTE pack_cur.proc4_6rows

Result of Executing PACK_CUR

The state of a package variable or cursor persists across transactions within a session. The state doesnot persist from session to session for the same user, nor does it persist from user to user.

Instructor Note

If you demonstrate this procedure, the results may vary because you may have added employees.

A more realistic way of sharing a result set between two procedures would be to use a cursorvariable as a parameter.


Passing Tables of Records to Procedures or Functions Inside a PackageInvoke the READ_EMP_TABLE procedure from an anonymous PL/SQL block, using iSQL*Plus.DECLARE

v_emp_table emp_package.emp_table_type;

BEGIN

emp_package.read_emp_table(v_emp_table);

DBMS_OUTPUT.PUT_LINE('An example: '||v_emp_table(4).last_name);

END;

/

To invoke the READ_EMP_TABLE procedure from another procedure or any PL/SQL block outsidethe package, the actual parameter referring to the OUT parameter P_EMP_TABLE must be prefixedwith its package name. In the example above, the V_EMP_TABLE variable is declared of theEMP_TABLE_TYPE type with the package name added as a prefix.

Instructor NoteIn the anonymous block above, EMP_PACKAGE.EMP_TABLE_TYPE refers to the user-definedpackaged table that is shown in the slide.


CREATE OR REPLACE PACKAGE BODY emp_package ISPROCEDURE read_emp_table

(p_emp_table OUT emp_table_type) ISi BINARY_INTEGER := 0;BEGIN

FOR emp_record IN (SELECT * FROM employees)LOOP

p_emp_table(i) := emp_record;i:= i+1;

END LOOP;END read_emp_table;

END emp_package;/

PL/SQL Tablesand Records in Packages

CREATE OR REPLACE PACKAGE emp_package ISTYPE emp_table_type IS TABLE OF employees%ROWTYPE

INDEX BY BINARY_INTEGER;PROCEDURE read_emp_table

(p_emp_table OUT emp_table_type);END emp_package;/



Summary


• Overload subprograms

• Use forward referencing

• Use one-time-only procedures

• Describe the purity level of package functions

• Identify the persistent state of packaged objects

Summary

Overloading is a feature that enables you to define different subprograms with the same name. It islogical to give two subprograms the same name in situations when the processing in both thesubprograms is the same, but the parameters passed to them varies.

PL/SQL allows for a special subprogram declaration called a forward declaration. Forwarddeclaration enables you to define subprograms in logical or alphabetical order, define mutuallyrecursive subprograms, and group subprograms in a package.

A one-time-only procedure is executed only when the package is first invoked within the other usersession. You can use this feature to initialize variables only once per session.

You can keep track of the state of a package variable or cursor, which persists throughout the usersession, from the time the user first references the variable or cursor to the time that the userdisconnects.





• Using overloaded subprograms

• Creating a one-time-only procedure


In this practice you create a package containing an overloaded function. You also create a one-time-only procedure within a package to populate a PL/SQL table.


Practice 131. Create a package called OVER_LOAD. Create two functions in this package; name each function

PRINT_IT. The function accepts a date or character string and prints a date or a number, dependingon how the function is invoked.

Note:• To print the date value, use DD-MON-YY as the input format, and FmMonth,dd yyyy as the

output format. Make sure you handle invalid input.• To print out the number, use 999,999.00 as the input format.

a. Test the first version of PRINT_IT with the following set of commands:

VARIABLE display_date VARCHAR2(20)

EXECUTE :display_date := over_load.print_it (TO_DATE('08-MAR-01'))

PRINT display_date

b. Test the second version of PRINT_IT with the following set of commands:

VARIABLE g_emp_sal NUMBER

EXECUTE :g_emp_sal := over_load.print_it('33,600')

PRINT g_emp_sal

2. Create a new package, called CHECK_PACK, to implement a new business rule.

a. Create a procedure called CHK_DEPT_JOB to verify whether a given combination ofdepartment ID and job is a valid one. In this case valid means that it must be a combination thatcurrently exists in the EMPLOYEES table.

Note:

• Use a PL/SQL table to store the valid department and job combination.

• The PL/SQL table needs to be populated only once.

• Raise an application error with an appropriate message if the combination is not valid.b. Test your CHK_DEPT_JOB package procedure by executing the following command:

EXECUTE check_pack.chk_dept_job(50,'ST_CLERK')

What happens?c. Test your CHK_DEPT_JOB package procedure by executing the following command:



Schedule: Timing Topic50 minutes Lecture

50 minutes Practice

100 minutes Total


Oracle Supplied Packages



Objectives

After completing this lesson, you should be able todo the following:• Write dynamic SQL statements using DBMS_SQL

and EXECUTE IMMEDIATE

• Describe the use and application of some Oracleserver-supplied packages:– DBMS_DDL

– DBMS_JOB

– DBMS_OUTPUT

– UTL_FILE

– UTL_HTTP and UTL_TCP

Lesson Aim

In this lesson, you learn how to use some of the Oracle server supplied packages and to take advantageof their capabilities.

Instructor Note

In this lesson, only a few of the many Oracle server supplied packages and their procedures, functions,and parameters are explained.


Using Supplied Packages

Packages are provided with the Oracle server to allow either PL/SQL access to certain SQL features, orto extend the functionality of the database.

You can take advantage of the functionality provided by these packages when creating your application,or you may simply want to use these packages as ideas when you create your own stored procedures.

Most of the standard packages are created by running catproc.sql.

Instructor NoteThe catproc.sql script is found in the $ORACLE_HOME/rdbms/admin directory. Otherpackages may have to be created in the SYS schema by running corresponding scripts located in thedirectory $ORACLE_HOME/rdbms/admin. The scripts to create supplied packages have prefixDBMS_.


Using Supplied Packages

Oracle-supplied packages:

• Are provided with the Oracle server

• Extend the functionality of the database

• Enable access to certain SQL features normallyrestricted for PL/SQL



Using Native Dynamic SQL

Dynamic SQL:

• Is a SQL statement that contains variables that canchange during runtime

• Is a SQL statement with placeholders and is storedas a character string

• Enables general-purpose code to be written

• Enables data-definition, data-control, or session-control statements to be written and executedfrom PL/SQL

• Is written using either DBMS_SQL or native dynamicSQL

Using Native Dynamic SQL (Dynamic SQL)

You can write PL/SQL blocks that use dynamic SQL. Dynamic SQL statements are not embedded inyour source program but rather are stored in character strings that are input to, or built by, the program.That is, the SQL statements can be created dynamically at run time by using variables. For example,you use dynamic SQL to create a procedure that operates on a table whose name is not known until runtime, or to write and execute a data definition language (DDL) statement (such as CREATE TABLE), adata control statement (such as GRANT), or a session control statement (such as ALTER SESSION). InPL/SQL, such statements cannot be executed statically.

In Oracle8, and earlier, you have to use DBMS_SQL to write dynamic SQL.

In Oracle 8i, you can use DBMS_SQL or native dynamic SQL. The EXECUTE IMMEDIATE statementcan perform dynamic single-row queries. Also, this is used for functionality such as objects andcollections, which are not supported by DBMS_SQL. If the statement is a multirow SELECT statement,you use OPEN-FOR, FETCH, and CLOSE statements.



Execution Flow

SQL statements go through various stages:

• Parse

• Bind

• Execute

• Fetch

Note: Some stages may be skipped.

Steps to Process SQL StatementsAll SQL statements have to go through various stages. Some stages may be skipped.

ParseEvery SQL statement must be parsed. Parsing the statement includes checking the statement's syntaxand validating the statement, ensuring that all references to objects are correct, and ensuring that therelevant privileges to those objects exist.

BindAfter parsing, the Oracle server knows the meaning of the Oracle statement but still may not haveenough information to execute the statement. The Oracle server may need values for any bind variablein the statement. The process of obtaining these values is called binding variables.

ExecuteAt this point, the Oracle server has all necessary information and resources, and the statement isexecuted.

FetchIn the fetch stage, rows are selected and ordered (if requested by the query), and each successive fetchretrieves another row of the result, until the last row has been fetched. You can fetch queries, but notthe DML statements.

Instructor NoteDo not go into too much detail when discussing this topic. Processing of SQL statements is covered inmore detail in other courses.



Using the DBMS_SQL Package

The DBMS_SQL package is used to write dynamic SQLin stored procedures and to parse DDL statements.Some of the procedures and functions of the packageinclude:

– OPEN_CURSOR

– PARSE

– BIND_VARIABLE

– EXECUTE

– FETCH_ROWS

– CLOSE_CURSOR

Using the DBMS_SQL Package

Using DBMS_SQL, you can write stored procedures and anonymous PL/SQL blocks that use dynamicSQL.

DBMS_SQL can issue data definition language statements in PL/SQL. For example, you can choose toissue a DROP TABLE statement from within a stored procedure.

The operations provided by this package are performed under the current user, not under the packageowner SYS. Therefore, if the caller is an anonymous PL/SQL block, the operations are performedaccording to the privileges of the current user; if the caller is a stored procedure, the operations areperformed according to the owner of the stored procedure.

Using this package to execute DDL statements can result in a deadlock. The most likely reason for thisis that the package is being used to drop a procedure that you are still using.


Components of the DBMS_SQL Package

The DBMS_SQL package uses dynamic SQL to access the database.

Instructor NoteDBMS_SQL has many more procedures and functions than are shown here. The procedures andfunctions in the table should show the main stages of dynamic SQL statements. If you want to coverDBMS_SQL in more detail, start by showing the data dictionary information on DBMS_SQL.SELECT text FROM all_source WHERE name ='DBMS_SQL' ORDER BY LINE;

orUse DESC DBMS_SQL to view the description of the procedures and functions in the package.The above SELECT statement returns 1050 rows.

Function or Procedure Description

OPEN_CURSOR Opens a new cursor and assigns a cursor ID number

PARSE Parses the DDL or DML statement: that is, checks the statement’s syntaxand associates it with the opened cursor (DDL statements are immediatelyexecuted when parsed)

BIND_VARIABLE Binds the given value to the variable identified by its name in the parsedstatement in the given cursor

EXECUTE Executes the SQL statement and returns the number of rows processed

FETCH_ROWS Retrieves a row for the specified cursor (for multiple rows, call in a loop)

CLOSE_CURSOR Closes the specified cursor


Example of a DBMS_SQL Package

In the slide, the table name is passed into the DELETE_ALL_ROWS procedure by using an INparameter. The procedure uses dynamic SQL to delete rows from the specified table. The numberof rows that are deleted as a result of the successful execution of the dynamic SQL are passed tothe calling environment through an OUT parameter.

How to Process Dynamic DML

1. Use OPEN_CURSOR to establish an area in memory to process a SQL statement.

2. Use PARSE to establish the validity of the SQL statement.

3. Use the EXECUTE function to run the SQL statement. This function returns the number ofrow processed.

4. Use CLOSE_CURSOR to close the cursor.

Instructor NoteIf you want to demonstrate this code, remember to do a ROLLBACK in order to undo the deletion.The sample code contains DELETE command on another test table that is created in the script.

Demonstration: 14_dmldynam.sql, 14_ddldynam.sql, and 14_seldynam.sql

Purpose: The first two scripts create a table to perform either a DROP TABLE or a DELETEcommand on it. The last script accepts three parameters and selects data from the table. UseEMPLOYEES, JOB, and MANAGER as the values when prompted. The values are not case-sensitive.


Using DBMS_SQL

Use dynamic SQL to delete rowsVARIABLE deleted NUMBEREXECUTE delete_all_rows('employees', :deleted)PRINT deleted

CREATE OR REPLACE PROCEDURE delete_all_rows(p_tab_name IN VARCHAR2, p_rows_del OUT NUMBER)

IScursor_name INTEGER;

BEGINcursor_name := DBMS_SQL.OPEN_CURSOR;DBMS_SQL.PARSE(cursor_name, 'DELETE FROM '||p_tab_name,

DBMS_SQL.NATIVE );p_rows_del := DBMS_SQL.EXECUTE (cursor_name);

DBMS_SQL.CLOSE_CURSOR(cursor_name);END;/



Use the EXECUTE IMMEDIATE statement for nativedynamic SQL with better performance.

• INTO is used for single-row queries and specifiesthe variables or records into which column valuesare retrieved.

• USING is used to hold all bind arguments. Thedefault parameter mode is IN.

Using the EXECUTE IMMEDIATE Statement

EXECUTE IMMEDIATE dynamic_string[INTO {define_variable

[, define_variable] ... | record}][USING [IN|OUT|IN OUT] bind_argument

[, [IN|OUT|IN OUT] bind_argument] ... ];

Using the EXECUTE IMMEDIATE Statement

Syntax Definition

You can use the INTO clause for a single-row query, but you must use OPEN-FOR, FETCH, andCLOSE for a multirow query.

Note: The syntax shown in the slide is not complete. The other clauses of the statement are discussed inthe Advanced PL/SQL course.

Instructor NoteTo process most dynamic SQL statements, you use the EXECUTE IMMEDIATE statement. However,to process a multirow query (SELECT statement), you must use the OPEN-FOR, FETCH, and CLOSEstatements.

OPEN-FOR, FETCH, and CLOSE are not covered in this course, because they use cursor variables.

Parameter Descriptiondynamic_string A string expression that represents a dynamic SQL statement (without

terminator) or a PL/SQL block (with terminator)define_variable A variable that stores the selected column valuerecord A user-defined or %ROWTYPE record that stores a selected rowbind_argument An expression whose value is passed to the dynamic SQL statement or

PL/SQL block


Using the EXECUTE IMMEDIATE Statement (continued)

In the EXECUTE IMMEDIATE statement:

• The INTO clause specifies the variables or record into which column values are retrieved. It isused only for single-row queries. For each value retrieved by the query, there must be acorresponding, type-compatible variable or field in the INTO clause.

• The RETURNING INTO clause specifies the variables into which column values are returned.It is used only for DML statements that have a RETURNING clause (without a BULKCOLLECT clause). For each value returned by the DML statement, there must be acorresponding, type-compatible variable in the RETURNING INTO clause.

• The USING clause holds all bind arguments. The default parameter mode is IN. For DMLstatements that have a RETURNING clause, you can place OUT arguments in the RETURNINGINTO clause without specifying the parameter mode, which, by definition, is OUT. If you useboth the USING clause and the RETURNING INTO clause, the USING clause can contain onlyIN arguments.

At run time, bind arguments replace corresponding placeholders in the dynamic string. Thus, everyplaceholder must be associated with a bind argument in the USING clause or RETURNING INTOclause. You can use numeric, character, and string literals as bind arguments, but you cannot useBoolean literals (TRUE, FALSE, and NULL).

Dynamic SQL supports all the SQL data types. For example, define variables and bind arguments canbe collections, LOBs, instances of an object type, and REFs. As a rule, dynamic SQL does not supportPL/SQL-specific types. For example, define variables and bind arguments cannot be Booleans orindex-by tables. The only exception is that a PL/SQL record can appear in the INTO clause.

You can execute a dynamic SQL statement repeatedly, using new values for the bind arguments.However, you incur some overhead because EXECUTE IMMEDIATE reparses the dynamic stringbefore every execution.



Dynamic SQL Using EXECUTE IMMEDIATE

CREATE PROCEDURE del_rows(p_table_name IN VARCHAR2,p_rows_deld OUT NUMBER)

ISBEGINEXECUTE IMMEDIATE 'delete from '||p_table_name;p_rows_deld := SQL%ROWCOUNT;

END;/

VARIABLE deleted NUMBEREXECUTE del_rows('test_employees',:deleted)PRINT deleted

Dynamic SQL Using EXECUTE IMMEDIATE

This is the same dynamic SQL as seen with DBMS_SQL, using the Oracle8i statement EXECUTEIMMEDIATE. The EXECUTE IMMEDIATE statement prepares (parses) and immediately executes thedynamic SQL statement.

Instructor NoteIf you run this example, remember to use ROLLBACK to undo the deletion.


Using the DBMS_DDL package

This package provides access to some SQL DDL statements, which you can use in PL/SQL programs.

DBMS_DDL is not allowed in triggers, in procedures called from Forms Builder, or in remote sessions.This package runs with the privileges of calling user, rather than the package owner SYS.

Practical Uses

• You can recompile your modified PL/SQL program units by usingDBMS_DDL.ALTER_COMPILE. The object type must be either procedure, function, package,package body, or trigger.

• You can analyze a single object, using DBMS_DDL.ANALYZE_OBJECT. (There is a way ofanalyzing more than one object at a time, using DBMS_UTILITY.) The object type should beTABLE, CLUSTER, or INDEX. The method must be COMPUTE, ESTIMATE, or DELETE.

• This package gives developers access to ALTER and ANALYZE SQL statements through PL/SQLenvironments.

Instructor NoteYou can check the column LAST_DDL_TIME in USER_OBJECTS and LAST_ANALYZED inUSER_TABLES to monitor the results of these operations.


Using the DBMS_DDL Package

The DBMS_DDL Package:

• Provides access to some SQL DDL statementsfrom stored procedures

• Includes some procedures:– ALTER_COMPILE (object_type, owner, object_name)

– ANALYZE_OBJECT (object_type, owner, name,method)

Note: This package runs with the privileges of callinguser, rather than the package owner SYS.

DBMS_DDL.ALTER_COMPILE('PROCEDURE','A_USER','QUERY_EMP')

DBMS_DDL.ANALYZE_OBJECT('TABLE','A_USER','JOBS','COMPUTE')



Using DBMS_JOB for Scheduling

DBMS_JOB Enables the scheduling and execution ofPL/SQL programs:

• Submitting jobs

• Executing jobs

• Changing execution parameters of jobs

• Removing jobs

• Suspending Jobs

Scheduling Jobs by Using DBMS_JOB

The package DBMS_JOB is used to schedule PL/SQL programs to run. Using DBMS_JOB, you cansubmit PL/SQL programs for execution, execute PL/SQL programs on a schedule, identify whenPL/SQL programs should run, remove PL/SQL programs from the schedule, and suspend PL/SQLprograms from running.

It can be used to schedule batch jobs during nonpeak hours or to run maintenance programs duringtimes of low usage.

Instructor NoteIn the init.ora file, the JOB_QUEUE_PROCESSES parameter must be set to greater than zero (atleast 1), because this parameter enables job queue processing in the background.



DBMS_JOB Subprograms

Available subprograms include:• SUBMIT

• REMOVE

• CHANGE

• WHAT

• NEXT_DATE

• INTERVAL

• BROKEN

• RUN

DBMS_JOB Subprograms

Subprogram DescriptionSUBMIT Submits a job to the job queueREMOVE Removes a specified job from the job queueCHANGE Alters a specified job that has already been submitted to the

job queue (you can alter the job description, the time atwhich the job will be run, or the interval between executionsof the job)

WHAT Alters the job description for a specified jobNEXT_DATE Alters the next execution time for a specified jobINTERVAL Alters the interval between executions for a specified jobBROKEN Disables job execution (if a job is marked as broken, the

Oracle server does not attempt to execute it)RUN Forces a specified job to run



Submitting Jobs

You can submit jobs by using DBMS_JOB.SUBMIT.

Available parameters include:• JOB OUT BINARY_INTEGER

• WHAT IN VARCHAR2

• NEXT_DATE IN DATE DEFAULT SYSDATE

• INTERVAL IN VARCHAR2 DEFAULT 'NULL'

• NO_PARSE IN BOOLEAN DEFAULT FALSE

DBMS_JOB.SUBMIT Parameters

The DBMS_JOB.SUBMIT procedure adds a new job to the job queue. It accepts five parameters andreturns the number of a job submitted through the OUT parameter JOB. The descriptions of theparameters are listed below.

Note: An exception is raised if the interval does not evaluate to a time in the future.

P a ra m eter M od e D escr ip t ionJOB OUT U nique identifier of the jo bWHAT IN PL /SQ L code to execu te as a jobNEXT_DATE IN N ex t ex ecu tio n da te o f the jobINTERVAL IN D ate fu nc tio n to com pu te the n ex t execu tio n da te o f a jobNO_PARSE IN B oo lean flag th a t ind icates w h ether to parse the jo b at jo b

su bm ission (the default is fa lse )



Submitting Jobs

Use DBMS_JOB.SUBMIT to place a job to be executedin the job queue.VARIABLE jobno NUMBERBEGIN

DBMS_JOB.SUBMIT (job => :jobno,what => 'OVER_PACK.ADD_DEPT(''EDUCATION'',2710);',next_date => TRUNC(SYSDATE + 1),interval => 'TRUNC(SYSDATE + 1)'

);COMMIT;

END;/PRINT jobno

ExampleThe block of code in the slide submits the ADD_DEPT procedure of the OVER_PACK package to thejob queue. The job number is returned through the JOB parameter. The WHAT parameter must beenclosed in single quotation marks and must include a semicolon at the end of the text string. This jobis submitted to run every day at midnight.

Note: In the example, the parameters are passed using named notation.

The transactions in the submitted job are not committed until either COMMIT is issued, orDBMS_JOB.RUN is executed to run the job. COMMIT in the slide commits the transaction.

Instructor NoteYou can demonstrate this code with the 14_16s.sql file.



Changing Job Characteristics

• DBMS_JOB.CHANGE: Changes the WHAT, NEXT_DATE,and INTERVAL parameters

• DBMS_JOB.INTERVAL: Changes the INTERVALparameter

• DBMS_JOB.NEXT_DATE: Changes the next executiondate

• DBMS_JOB.WHAT: Changes the WHAT parameter

Changing Jobs After Being SubmittedThe CHANGE, INTERVAL, NEXT_DATE, and WHAT procedures enable you to modify jobcharacteristics after a job is submitted to the queue. Each of these procedures takes the JOB parameteras an IN parameter indicating which job is to be changed.

Example

The following code changes job number 1 to execute on the following day at 6:00 a.m. and every fourhours after that.BEGIN

DBMS_JOB.CHANGE(1, NULL, TRUNC(SYSDATE+1)+6/24, ’SYSDATE+4/24');END;/

Note: Each of these procedures can be executed on jobs owned by the username to which the sessionis connected. If the parameter what, next_date, or interval is NULL, then the last valuesassigned to those parameters are used.

Instructor Note

You can demonstrate this code with the 14_17n.sql file.



Running, Removing, and Breaking Jobs

• DBMS_JOB.RUN: Runs a submitted job immediately

• DBMS_JOB.REMOVE: Removes a submitted job fromthe job queue

• DBMS_JOB.BROKEN: Marks a submitted job asbroken, and a broken job will not run

Running, Removing, and Breaking JobsThe DBMS_JOB.RUN procedure executes a job immediately. Pass the job number that you want torun immediately to the procedure.

EXECUTE DBMS_JOB.RUN(1)

The DBMS_JOB.REMOVE procedure removes a submitted job from the job queue. Pass the jobnumber that you want to remove from the queue to the procedure.

EXECUTE DBMS_JOB.REMOVE(1)

The DBMS_JOB.BROKEN marks a job as broken or not broken. Jobs are not broken by default. Youcan change a job to the broken status. A broken job will not run. There are three parameters for thisprocedure. The JOB parameter identifies the job to be marked as broken or not broken. The BROKENparameter is a Boolean parameter. Set this parameter to FALSE to indicate that a job is not broken,and set it to TRUE to indicate that it is broken. The NEXT_DATE parameter identifies the nextexecution date of the job.

EXECUTE DBMS_JOB.BROKEN(1, TRUE)



Viewing Information on Submitted Jobs

• Use the DBA_JOBS dictionary view to see thestatus of submitted jobs.

• Use the DBA_JOBS_RUNNING dictionary view todisplay jobs that are currently running.

SELECT job, log_user, next_date, next_sec,broken, what

FROM DBA_JOBS;

Viewing Information on Submitted JobsThe DBA_JOBS and DBA_JOBS_RUNNING dictionary views display information about jobs in thequeue and jobs that have run. To be able to view the dictionary information, users should be grantedthe SELECT privilege on SYS.DBA_JOBS.

The query shown in the slide displays the job number, the user who submitted the job, the scheduleddate for the job to run, the time for the job to run, and the PL/SQL block executed as a job.

Use the USER_JOBS data dictionary view to display information about jobs in the queue for you.This view has the same structure as the DBA_JOBS view.

Instructor Note

You may want to describe these views in iSQL*Plus. The columns are fairly descriptive of theinformation held in them. Refer students to the reference material for more in-depth descriptions.

You can demonstrate this code with the 14_19s.sql file.


Using the DBMS_OUTPUT Package

The DBMS_OUTPUT package outputs values and messages from any PL/SQL block.

Practical Uses

• You can output intermediary results to the window for debugging purposes.

• This package enables developers to closely follow the execution of a function or procedure bysending messages and values to the output buffer.

Function or Procedure Description

PUT Appends text from the procedure to the current line of the lineoutput buffer

NEW_LINE Places an end_of_line marker in the output buffer

PUT_LINE Combines the action of PUT and NEW_LINE

GET_LINE Retrieves the current line from the output buffer into theprocedure

GET_LINES Retrieves an array of lines from the output buffer into theprocedure

ENABLE/DISABLE Enables or disables calls to the DBMS_OUTPUT procedures


Using the DBMS_OUTPUT Package

The DBMS_OUTPUT package enables you to outputmessages from PL/SQL blocks. Available proceduresinclude:• PUT

• NEW_LINE

• PUT_LINE

• GET_LINE

• GET_LINES

• ENABLE/DISABLE



Interacting with Operating System Files

• UTL_FILE Oracle-supplied package:

– Provides text file I/O capabilities

– Is available with version 7.3 and later

• The DBMS_LOB Oracle-supplied package:– Provides read-only operations on external BFILES

– Is available with version 8 and later– Enables read and write operations on internal LOBs

Interacting with Operating System Files

Two Oracle-supplied packages are provided. You can use them to access operating system files.

With the Oracle-supplied UTL_FILE package, you can read from and write to operating system files.This package is available with database version 7.3 and later and the PL/SQL version 2.3 and later.

With the Oracle-supplied package DBMS_LOB, you can read from binary files on the operatingsystem. This package is available from the database version 8.0 and later. This package is discussedlater in the lesson “Manipulating Large Objects.”



What Is the UTL_FILE Package?

• Extends I/O to text files within PL/SQL

• Provides security for directories on the serverthrough the init.ora file

• Is similar to standard operating system I/O– Open files

– Get text

– Put text

– Close files– Use the exceptions specific to the UTL_FILE

package

The UTL_FILE PackageThe UTL_FILE package provides text file I/O from within PL/SQL. Client-side securityimplementation uses normal operating system file permission checking. Server-side security isimplemented through restrictions on the directories that can be accessed. In the init.ora file, theinitialization parameter UTL_FILE_DIR is set to the accessible directories desired.

UTL_FILE_DIR = directory_name

For example, the following initialization setting indicates that the directory/usr/ngreenbe/my_app is accessible to the fopen function, assuming that the directory isaccessible to the database server processes. This parameter setting is case-sensitive on case-sensitive operating systems.

UTL_FILE_DIR = /user/ngreenbe/my_app

The directory should be on the same machine as the database server. Using the following settingturns off database permissions and makes all directories that are accessible to the database serverprocesses also accessible to the UTL_FILE package.

UTL_FILE_DIR = *

Using the procedures and functions in the package, you can open files, get text from files, put textinto files, and close files. There are seven exceptions declared in the package to account for possibleerrors raised during execution.

Instructor NoteUTL_FILE_DIR = * is not recommended. Your init.ora file needs to include the parameterUTL_FILE_DIR. If it does not, edit the init.ora file and add the parameter. Then restart thedatabase by using the SHUTDOWN and STARTUP commands.



YesGet linesfrom thetext file

Put linesinto thetext file

Morelines toprocess?

No Closethe

text file

File Processing Using theUTL_FILE Package

Open thetext file

File Processing Using the UTL_FILE Package

Before using the UTL_FILE package to read from or write to a text file, you must first check whetherthe text file is open by using the IS_OPEN function. If the file is not open, you open the file with theFOPEN function. You then either read the file or write to the file until processing is done. At the endof file processing, use the FCLOSE procedure to close the file.

Note: A summary of the procedures and functions within the UTL_FILE package is listed on the nextpage.

Instructor Note

Reading or writing to a file in PL/SQL or SQL is the same as reading or writing to files in other third-generation languages.



UTL_FILE Procedures and Functions

• Function FOPEN

• Function IS_OPEN

• Procedure GET_LINE

• Procedure PUT, PUT_LINE, PUTF

• Procedure NEW_LINE

• Procedure FFLUSH

• Procedure FCLOSE, FCLOSE_ALL

The UTL_FILE Package: Procedures and Functions

Note: The maximum size of an input record is 1,023 bytes unless you specify a larger size in theoverloaded version of FOPEN.

Function or Procedure DescriptionFOPEN A function that opens a file for input or output and returns a file

handle used in subsequent I/O operationsIS_OPEN A function that returns a Boolean value whenever a file handle

refers to an open fileGET_LINE A procedure that reads a line of text from the opened file and

places the text in the output buffer parameter (the maximum sizeof an input record is 1,023 bytes unless you specify a larger sizein the overloaded version of FOPEN)

PUT, PUT_LINE A procedure that writes a text string stored in the bufferparameter to the opened file (no line terminator is appended byput; use new_line to terminate the line, or use PUT_LINEto write a complete line with a terminator)

PUTF A formatted put procedure with two format specifiers: %s and\n (use %s to substitute a value into the output string. \n is anew line character)

NEW_LINE Procedure that terminates a line in an output file

FFLUSH Procedure that writes all data buffered in memory to a fileFCLOSE Procedure that closes an opened fileFCLOSE_ALL Procedure that closes all opened file handles for the session



Exceptions Specific to the UTL_FILEPackage

• INVALID_PATH

• INVALID_MODE

• INVALID_FILEHANDLE

• INVALID_OPERATION

• READ_ERROR

• WRITE_ERROR

• INTERNAL_ERROR

Exceptions to the UTL_FILE Package

The UTL_FILE package declares seven exceptions that are raised to indicate an error condition in theoperating system file processing.

Note: These exceptions must be prefaced with the package name.

UTL_FILE procedures can also raise predefined PL/SQL exceptions such as NO_DATA_FOUND orVALUE_ERROR.

Exception Name DescriptionINVALID_PATH The file location or filename was invalid.

INVALID_MODE The OPEN_MODE parameter in FOPEN was invalid.

INVALID_FILEHANDLE The file handle was invalid.

INVALID_OPERATION The file could not be opened or operated on as requested.

READ_ERROR An operating system error occurred during the read operation.

WRITE_ERROR An operating system error occurred during the write operation.

INTERNAL_ERROR An unspecified error occurred in PL/SQL.



FUNCTION FOPEN

(location IN VARCHAR2,

filename IN VARCHAR2,

open_mode IN VARCHAR2)

RETURN UTL_FILE.FILE_TYPE;

FUNCTION IS_OPEN

(file_handle IN FILE_TYPE)

RETURN BOOLEAN;

The FOPEN and IS_OPEN Functions

FOPEN Function Parameters

Syntax Definitions

The return value is the file handle that is passed to all subsequent routines that operate on the file.

IS_OPEN Function

The function IS_OPEN tests a file handle to see if it identifies an opened file. It returns a Booleanvalue indicating whether the file has been opened but not yet closed.

Note: For the full syntax, refer to Oracle9i Supplied PL/SQL Packages and Types Reference.

Where location Is the operating-system-specific string thatspecifies the directory or area in which toopen the file

filename Is the name of the file, including theextension, without any pathing information

open_mode Is string that specifies how the file is to beopened; Supported values are:‘r’ read text (use GET_LINE)‘w’ write text (PUT, PUT_LINE,

NEW_LINE, PUTF,FFLUSH)

‘a’ append text (PUT, PUT_LINE,NEW_LINE, PUTF,FFLUSH)



CREATE OR REPLACE PROCEDURE sal_status

(p_filedir IN VARCHAR2, p_filename IN VARCHAR2)

IS

v_filehandle UTL_FILE.FILE_TYPE;

CURSOR emp_info IS

SELECT last_name, salary, department_id

FROM employees

ORDER BY department_id;

v_newdeptno employees.department_id%TYPE;

v_olddeptno employees.department_id%TYPE := 0;

BEGIN

v_filehandle := UTL_FILE.FOPEN (p_filedir, p_filename,'w');

UTL_FILE.PUTF (v_filehandle,'SALARY REPORT: GENERATED ON

%s\n', SYSDATE);

UTL_FILE.NEW_LINE (v_filehandle);

FOR v_emp_rec IN emp_info LOOP

v_newdeptno := v_emp_rec.department_id;

...

Using UTL_FILE

sal_status.sql

Using UTL_FILE

ExampleThe SAL_STATUS procedure creates a report of employees for each department and their salaries.This information is sent to a text file by using the UTL_FILE procedures and functions.

The variable v_filehandle uses a type defined in the UTL_FILE package. This package definedtype is a record with a field called ID of the BINARY_INTEGER datatype.

TYPE file_type IS RECORD (id BINARY_INTEGER);

The contents of file_type are private to the UTL_FILE package. Users of the package should notreference or change components of this record.

The names of the text file and the location for the text file are provided as parameters to the program.EXECUTE sal_status('C:\UTL_FILE', 'SAL_RPT.TXT')

Note: The file location shown in the above example is defined as value of UTL_FILE_DIR in theinit.ora file as follows: UTL_FILE_DIR = C:\UTL_FILE.When reading a complete file in a loop, you need to exit the loop using the NO_DATA_FOUNDexception. UTL_FILE output is sent synchronously. DBMS_OUTPUT procedures do not produceoutput until the procedure is completed.

Instructor NoteYou can demonstrate this code with the 14_27s.sql and 14_27n.sql files. Run the script14_27s.sql first. You can demonstrate the output generated in the file instructor.txt afteryou run 14_27n.sql.



...

IF v_newdeptno <> v_olddeptno THEN

UTL_FILE.PUTF (v_filehandle, 'DEPARTMENT: %s\n',

v_emp_rec.department_id);

END IF;

UTL_FILE.PUTF (v_filehandle,' EMPLOYEE: %s earns: %s\n',

v_emp_rec.last_name, v_emp_rec.salary);

v_olddeptno := v_newdeptno;

END LOOP;

UTL_FILE.PUT_LINE (v_filehandle, '*** END OF REPORT ***');

UTL_FILE.FCLOSE (v_filehandle);

EXCEPTION

WHEN UTL_FILE.INVALID_FILEHANDLE THEN

RAISE_APPLICATION_ERROR (-20001, 'Invalid File.');

WHEN UTL_FILE.WRITE_ERROR THEN

RAISE_APPLICATION_ERROR (-20002, 'Unable to write tofile');

END sal_status;/

Using UTL_FILE

sal_status.sql

Using UTL_FILE (continued)

The output for this report in the sal_rpt.txt file is as follows:

SALARY REPORT: GENERATED ON 08-MAR-01

DEPARTMENT: 10EMPLOYEE: Whalen earns: 4400

DEPARTMENT: 20EMPLOYEE: Hartstein earns: 13000EMPLOYEE: Fay earns: 6000

DEPARTMENT: 30EMPLOYEE: Raphaely earns: 11000EMPLOYEE: Khoo earns: 3100...

DEPARTMENT: 100EMPLOYEE: Greenberg earns: 12000...

DEPARTMENT: 110EMPLOYEE: Higgins earns: 12000EMPLOYEE: Gietz earns: 8300EMPLOYEE: Grant earns: 7000

*** END OF REPORT ***



The UTL_HTTP Package

The UTL_HTTP package:

• Enables HTTP callouts from PL/SQL and SQL toaccess data on the Internet

• Contains the functions REQUEST andREQUEST_PIECES which take the URL of a site as aparameter, contact that site, and return the dataobtained from that site

• Requires a proxy parameter to be specified in theabove functions, if the client is behind a firewall

• Raises INIT_FAILED or REQUEST_FAILEDexceptions if HTTP call fails

• Reports an HTML error message if specified URLis not accessible

The UTL_HTTP Package

UTL_HTTP is a package that allows you to make HTTP requests directly from the database. TheUTL_HTTP package makes hypertext transfer protocol (HTTP) callouts from PL/SQL and SQL. Youcan use it to access data on the Internet or to call Oracle Web Server Cartridges. By couplingUTL_HTTP with the DBMS_JOBS package, you can easily schedule reoccurring requests be made fromyour database server out to the Web.

This package contains two entry point functions: REQUEST and REQUEST_PIECES. Both functionstake a string universal resource locator (URL) as a parameter, contact the site, and return the HTML dataobtained from the site. The REQUEST function returns up to the first 2000 bytes of data retrieved fromthe given URL. The REQUEST_PIECES function returns a PL/SQL table of 2000-byte pieces of thedata retrieved from the given URL.

If the HTTP call fails, for a reason such as that the URL is not properly specified in the HTTP syntaxthen the REQUEST_FAILED exception is raised. If initialization of the HTTP-callout subsystem fails,for a reason such as a lack of available memory, then the INIT_FAILED exception is raised.

If there is no response from the specified URL, then a formatted HTML error message may be returned.

If REQUEST or REQUEST_PIECES fails by returning either an exception or an error message, thenverify the URL with a browser, to verify network availability from your machine. If you are behind afirewall, then you need to specify proxy as a parameter, in addition to the URL.

This package is covered in more detail in the course Administering Oracle9i Application Server.

For more information, refer to Oracle9i Supplied PL/SQL Packages Reference.



SELECT UTL_HTTP.REQUEST('http://www.oracle.com','edu-proxy.us.oracle.com')

FROM DUAL;

Using the UTL_HTTP Package

Using the UTL_HTTP PackageThe SELECT statement and the output in the slide show how to use the REQUEST function of theUTL_HTTP package to retrieve contents from the URL www.oracle.com. The second parameter to thefunction indicates the proxy because the client being tested is behind a firewall.

The retrieved output is in HTML format.

You can use the function in a PL/SQL block as shown below. The function retrieves up to 100 pieces ofdata, each of a maximum 2000 bytes from the URL. The number of pieces and the total length of the dataretrieved are printed.DECLAREx UTL_HTTP.HTML_PIECES;

BEGINx := UTL_HTTP.REQUEST_PIECES('http://www.oracle.com/',100,

'edu-proxy.us.oracle.com');DBMS_OUTPUT.PUT_LINE(x.COUNT || ' pieces were retrieved.');DBMS_OUTPUT.PUT_LINE('with total length ');IF x.COUNT < 1 THEN DBMS_OUTPUT.PUT_LINE('0');ELSE DBMS_OUTPUT.PUT_LINE((2000*(x.COUNT - 1))+LENGTH(x(x.COUNT)));END IF;

END;/



Using the UTL_TCP Package

The UTL_TCP Package:

• Enables PL/SQL applications to communicate withexternal TCP/IP-based servers using TCP/IP

• Contains functions to open and close connections,to read or write binary or text data to or from aservice on an open connection

• Requires remote host and port as well as local hostand port as arguments to its functions

• Raises exceptions if the buffer size is too small,when no more data is available to read from aconnection, when a generic network error occurs, orwhen bad arguments are passed to a function call

Using the UTL_TCP Package

The UTL_TCP package enables PL/SQL applications to communicate with external TCP/IP-based serversusing TCP/IP. Because many Internet application protocols are based on TCP/IP, this package is useful toPL/SQL applications that use Internet protocols.

The package contains functions such as:

OPEN_CONNECTION: This function opens a TCP/IP connection with the specified remote and local hostand port details. The remote host is the host providing the service. The remote port is the port number onwhich the service is listening for connections. The local host and port numbers represent those of the hostproviding the service. The function returns a connection of PL/SQL record type.

CLOSE_CONNECTION: This procedure closes an open TCP/IP connection. It takes the connection detailsof a previously opened connection as parameter. The procedure CLOSE_ALL_CONNECTIONS closes allopen connections.

READ_BINARY()/TEXT()/LINE(): This function receives binary, text, or text line data from aservice on an open connection.

WRITE_BINARY()/TEXT()/LINE(): This function transmits binary, text, or text line message to aservice on an open connection.

Exceptions are raised when buffer size for the input is too small, when generic network error occurs, whenno more data is available to read from the connection, or when bad arguments are passed in a function call.

This package is discussed in detail in the course Administering Oracle9i Application Server. For moreinformation, refer to Oracle 9i Supplied PL/SQL Packages Reference.


Using Oracle-Supplied Packages

Package Description

DBMS_ALERT Provides notification of database events

DBMS_APPLICATION_INFO Allows application tools and application developers to inform thedatabase of the high level of actions they are currently performing

DBMS_DESCRIBE Returns a description of the arguments for a stored procedure

DBMS_LOCK Requests, converts, and releases userlocks, which are managed bythe RDBMS lock management services

DBMS_SESSION Provides access to SQL session information

DBMS_SHARED_POOL Keeps objects in shared memory

DBMS_TRANSACTION Controls logical transactions and improves the performance ofshort, nondistributed transactions

DBMS_UTILITY Analyzes objects in a particular schema, checks whether the serveris running in parallel mode, and returns the time


Oracle-Supplied Packages

• DBMS_ALERT

• DBMS_APPLICATION_INFO

• DBMS_DESCRIBE

• DBMS_LOCK

• DBMS_SESSION

Other Oracle-supplied packages include:

• DBMS_SHARED_POOL

• DBMS_TRANSACTION

• DBMS_UTILITY


Built-in Name DescriptionCALENDAR Provides calendar maintenance functionsDBMS_ALERT Supports asynchronous notification of database events.

Messages or alerts are sent on a COMMIT command.Message transmittal is one way, but one sender can alertseveral receivers.

DBMS_APPLICATION_INFO Is used to register an application name with the databasefor auditing or performance tracking purposes

DBMS_AQ Provides message queuing as part of the Oracle server; isused to add a message (of a predefined object type) onto aqueue or dequeue a message

DBMS_AQADM Is used to perform administrative functions on a queue orqueue table for messages of a predefined object type

DBMS_DDL Is used to embed the equivalent of the SQL commandsALTER, COMPILE, and ANALYZE within your PL/SQLprograms

DBMS_DEBUG A PL/SQL API to the PL/SQL debugger layer, Probe, inthe Oracle server

DBSM_DEFERDBMS_DEFER_QUERYDBMS_DEFER_SYS

Is used to build and administer deferred remote procedurecalls (use of this feature requires the Replication Option)

DBMS_DESCRIBE Is used to describe the arguments of a stored procedureDBMS_DISTRIBRUTED_TRUST_ADMIN

Is used to maintain the Trusted Servers list, which is usedin conjunction with the list at the central authority todetermine whether a privileged database link from aparticular server can be accepted

DBMS_HS Is used to administer heterogeneous services byregistering or dropping distributed external procedures,remote libraries, and non-Oracle systems (you usedbms_hs to create or drop some initialization variablesfor non-Oracle systems)

DBMS_HS_EXTPROC Enables heterogeneous services to establish security fordistributed external procedures

DBMS_HS_PASSTHROUGH Enables heterogeneous services to send pass-through SQLstatements to non-Oracle systems

DBMS_IOT Is used to schedule administrative procedures that youwant performed at periodic intervals; is also the interfacefor the job queue

DBMS_JOB Is used to schedule administrative procedures that youwant performed at periodic intervals

DBMS_LOB Provides general purpose routines for operations onOracle large objects (LOBs) data types: BLOB, CLOB(read only) and BFILES (read-only)

Oracle-Supplied Packages

The following list summarizes and provides a brief description of the packages supplied with Oracle9i.


Built-in Nam e DescriptionDBMS_LOCK Is used to request, convert, and release locks through

Oracle Lock M anagement servicesDBMS_LOGMNR Provides functions to initialize and run the log readerDBMS_LOGMNR_D Queries the dictionary tables of the current database, and

creates a text based file containing their contentsDBMS_OFFLINE_OG Provides public APIs for offline instantiation of m aster

groupsDBMS_OFFLINE_SNAPSHOT

Provides public APIs for offline instantiation of snapshots

DBMS_OLAP Provides procedures for summaries, dim ensions, andquery rewrites

DBMS_ORACLE_TRACE_AGENT

Provides client callable interfaces to the Oracle TRAC Einstrumentation within the Oracle7 server

DBMS_ORACLE_TRACE_USER

Provides public access to the Oracle7 release serverOracle TR ACE instrum entation for the calling user

DBMS_OUTPUT Accum ulates information in a buffer so that it can beretrieved out later

DBMS_PCLXUTIL Provides intrapartition parallelism for creating partition-wise local indexes

DBMS_PIPE Provides a DBM S pipe service that enables messages tobe sent between sessions

DBMS_PROFILER Provides a Probe Profiler API to profile existing PL/SQLapplications and identify perform ance bottlenecks

DBMS_RANDOM Provides a built-in random num ber generatorDBMS_RECTIFIER_DIFF Provides APIs used to detect and resolve data

inconsistencies between two replicated sitesDBMS_REFRESH Is used to create groups of snapshots that can be refreshed

together to a transactionally consistent point in tim e;requires the Distributed option

DBMS_REPAIR Provides data corruption repair proceduresDBMS_REPCAT Provides routines to administer and update the replication

catalog and environment; requires the Replication optionDBMS_REPCAT_ADMIN Is used to create users with the privileges needed by the

symmetric replication facility; requires the Replicationoption

DBMS_REPCAT_INSTATIATE

Instantiates deployment tem plates; requires theR eplication option

DBMS_REPCAT_RGT C ontrols the maintenance and definition of refresh grouptemplates; requires the R eplication option

DBMS_REPUTIL Provides routines to generate shadow tables, triggers, andpackages for table replication

DBMS_RESOURCE_MANAGER

M aintains plans, consumer groups, and plan directives; italso provides semantics so that you may group togetherchanges to the plan schema

Oracle Supplied Packages (continued)


Built-in Name DescriptionDBMS_RESOURCE_MANAGER_PRIVS

Maintains privileges associated with resource consumergroups

DBMS_RLS Provides row-level security administrative interfaceDBMS_ROWID Is used to get information about ROWIDs, including the

data block number, the object number, and othercomponents

DBMS_SESSION Enables programmatic use of the SQL ALTER SESSIONstatement as well as other session-level commands

DBMS_SHARED_POOL Is used to keep objects in shared memory, so that they arenot aged out with the normal LRU mechanism

DBMS_SNAPSHOT Is used to refresh one or more snapshots that are not partof the same refresh group and purge logs; use of thisfeature requires the Distributed option

DBMS_SPACE Provides segment space information not available throughstandard views

DBMS_SPACE_ADMIN Provides tablespace and segment space administration notavailable through standard SQL

DSMS_SQL Is used to write stored procedure and anonymous PL/SQLblocks using dynamic SQL; also used to parse any DMLor DDL statement

DBMS_STANDARD Provides language facilities that help your applicationinteract with the Oracle server

DBMS_STATS Provides a mechanism for users to view and modifyoptimizer statistics gathered for database objects

DBMS_TRACE Provides routines to start and stop PL/SQL tracingDBMS_TRANSACTION Provides procedures for a programmatic interface to

transaction managementDBMS_TTS Checks whether if the transportable set is self-containedDBMS_UTILITY Provides functionality for managing procedures, reporting

errors, and other informationDEBUG_EXTPROC Is used to debug external procedures on platforms with

debuggers that can attach to a running processOUTLN_PKG Provides the interface for procedures and functions

associated with management of stored outlinesPLITBLM Handles index-table operationsSDO_ADMIN Provides functions implementing spatial index creation

and maintenance for spatial objectsSDO_GEOM Provides functions implementing geometric operations on

spatial objectsSDO_MIGRATE Provides functions for migrating spatial data from release

7.3.3 and 7.3.4 to 8.1.xSDO_TUNE Provides functions for selecting parameters that determine

the behavior of the spatial indexing scheme used in theSpatial Cartridge




Built-in Name DescriptionSTANDARD Declares types, exceptions, and subprograms that are

available automatically to every PL/SQL programTIMESERIES Provides functions that perform operations, such as

extraction, retrieval, arithmetic, and aggregation, on timeseries data

TIMESCALE Provides scale-up and scale-down functionsTSTOOLS Provides administrative tools proceduresUTL_COLL Enables PL/SQL programs to use collection locators to

query and updateUTL_FILE Enables your PL/SQL programs to read and write

operating system (OS) text files and provides a restrictedversion of standard OS stream file I/O

UTL_HTTP Enables HTTP callouts from PL/SQL and SQL to accessdata on the Internet or to call Oracle Web ServerCartridges

UTL_PG Provides functions for converting COBOL numeric datainto Oracle numbers and Oracle numbers into COBOLnumeric data

UTL_RAW Provides SQL functions for RAW data types thatconcatenate, obtain substring, and so on, to and from RAWdata types

UTL_REF Enables a PL/SQL program to access an object byproviding a reference to the object

VIR_PKG Provides analytical and conversion functions for visualinformation retrieval



Summary


• Take advantage of the preconfigured packagesthat are provided by Oracle

• Create packages by using the catproc.sql script

• Create packages individually.

DBMS Packages and the Scripts to Execute Them

Note: For more information about these packages and scripts, refer to Oracle9i Supplied PL/SQLPackages and Types Reference.

Instructor Note

Point out to the students that these script files often have useful comments embedded within them thatsupplement the documentation. See the dbmspipe.sql script for an example.

DBMS_ALERT dbmsalrt.sql

DBMS_APPLICATION_INFO dbmsutil.sql

DBMS_DDL dbmsutil.sql

DBMS_LOCK dbmslock.sql

DBMS_OUTPUT dbmsotpt.sql

DBMS_PIPE dbmspipe.sql

DBMS_SESSION dbmsutil.sql

DBMS_SHARED_POOL dbmsspool.sql

DBMS_SQL dbmssql.sql

DBMS_TRANSACTION dbmsutil.sql

DBMS_UTILITY dbmsutil.sql




This practice covers using:• DBMS_SQL for dynamic SQL

• DBMS_DDL to analyze a table

• DBMS_JOB to schedule a task

• UTL_FILE to generate text reports

Practice 14 OverviewIn this practice, you use DBMS_SQL to implement a procedure to drop a table. You also use theEXECUTE IMMEDIATE command to drop a table. You use DBMS_DDL to analyze objects in yourschema, and you can schedule the analyze procedure through DBMS_JOB.

In this practice, you also write a PL/SQL program that generates customer statuses into a text file.


Practice 141. a. Create a DROP_TABLE procedure that drops the table specified in the input parameter.

Use the procedures and functions from the supplied DBMS_SQL package.

b. To test the DROP_TABLE procedure, first create a new table called EMP_DUP as acopy of the EMPLOYEES table.

c. Execute the DROP_TABLE procedure to drop the EMP_DUP table.

2. a. Create another procedure called DROP_TABLE2 that drops the table specified in theinput parameter. Use the EXECUTE IMMEDIATE statement.

b. Repeat the test outlined in steps 1-b and 1-c.

3. a. Create a procedure called ANALYZE_OBJECT that analyzes the given object that youspecified in the input parameters. Use the DBMS_DDL package, and use the COMPUTEmethod.

b. Test the procedure using the EMPLOYEES table. Confirm that theANALYZE_OBJECT procedure has run by querying the LAST_ANALYZED column inthe USER_TABLES data dictionary view.

If you have time:

4. a. Schedule ANALYZE_OBJECT by using DBMS_JOB. Analyze the DEPARTMENTStable, and schedule the job to run in five minutes time from now. (To start the job infive minutes from now, set the parameter NEXT_DATE = 5/(24*60) = 1/288.)

b. Confirm that the job has been scheduled by using USER_JOBS.

5. Create a procedure called CROSS_AVGSAL that generates a text file report ofemployees who have exceeded the average salary of their department. The partialcode is provided for you in the file lab14_5.sql.

a. Your program should accept two parameters. The first parameter identifies the outputdirectory. The second parameter identifies the text file name to which your procedurewrites.

b. Your instructor will inform you of the directory location. When you invoke theprogram, name the second parameter sal_rptxx.txt where xx stands for youruser number, such as 01, 15, and so on.

c. Add an exception handling section to handle errors that may be encountered fromusing the UTL_FILE package.

Sample output from this file follows:EMPLOYEES OVER THE AVERAGE SALARY OF THEIR DEPARTMENT:REPORT GENERATED ON 26-FEB-01

Hartstein 20 $13,000.00Raphaely 30 $11,000.00Marvis 40 $6,500.00...*** END OF REPORT ***


Instructor NoteYou can locate the directory name UTL_FILE, where the students write their files as follows:

Use Telnet to connect to the server provided by the ESS group. Log in to the server operating systemusing the username and password supplied by ESS.

Type the command ls to list the contents and observe that the directory UTL_FILE is listed.

Type the command pwd to display the present working directory. Give this value to the students touse as a parameter in their practice.


45 minutes Practice

115 minutes Total


Manipulating Large Objects


Lesson Aim

Databases have long been used to store large objects. However, the mechanisms built into databaseshave never been as useful as the new large object (LOB) data types provided in Oracle8. This lessondescribes the characteristics of the new data types, comparing and contrasting them with earlier datatypes. Examples, syntax, and issues regarding the LOB types are also presented.

Note: A LOB is a data type and should not be confused with an object type.

Instructor NoteLOBs are available with Oracle8. They are not part of any option.


Objectives

After completing this lesson, you should be able todo the following:• Compare and contrast LONG and large object (LOB)

data types• Create and maintain LOB data types

• Differentiate between internal and external LOBs

• Use the DBMS_LOB PL/SQL package

• Describe the use of temporary LOBs



What Is a LOB?

Movie(BFILE)

Photo(BLOB)

“Four score and seven years ago

our fathers brought forth upon

this continent, a new nation,

conceived in LIBERTY, and dedicated

to the proposition that all men

are created equal.”

Text(CLOB)

LOBs are used to store large unstructured data such astext, graphic images, films, and sound waveforms.

OverviewA LOB is a data type that is used to store large, unstructured data such as text, graphic images, video clippings,and so on. Structured data such as a customer record may be a few hundred bytes, but even small amounts ofmultimedia data can be thousands of times larger. Also, multimedia data may reside on operating system (OS)files, which may need to be accessed from a database.There are four large object data types:• BLOB represents a binary large object, such as a video clip.

• CLOB represents a character large object.

• NCLOB represents a multibyte character large object.

• BFILE represents a binary file stored in an operating system binary file outside the database. The BFILEcolumn or attribute stores a file locator that points to the external file.

• LOBs are characterized in two ways, according to their interpretation by the Oracle server (binary orcharacter) and their storage aspects. LOBs can be stored internally (inside the database) or in host files.There are two categories of LOBs:

• Internal LOBs (CLOB, NCLOB, BLOB) are stored in the database.

• External files (BFILE) are stored outside the database.

The Oracle9i Server performs implicit conversion between CLOB and VARCHAR2 data types. The other implicitconversions between LOBs are not possible. For example, if the user creates a table T with a CLOB column anda table S with a BLOB column, the data is not directly transferable between these two columns.BFILEs can be accessed only in read-only mode from an Oracle server.


LONG and LOB Data Types

LONG and LONG RAW data types were previously used for unstructured data, such as binary images,documents, or geographical information. These data types are superseded by the LOB data types. Oracle9i provides a LONG-to-LOB API to migrate from LONG columns to LOB columns.

It is beneficial to discuss LOB functionality in comparison to the older types. In the bulleted list below,LONGs refers to LONG and LONG RAW, and LOBs refers to all LOB data types:

• A table can have multiple LOB columns and object type attributes. A table can have only one LONGcolumn.

• The maximum size of LONGs is 2 gigabytes; LOBs can be up to 4 gigabytes.

• LOBs return the locator; LONGs return the data.

• LOBs store a locator in the table and the data in a different segment, unless the data is less than4,000 bytes; LONGs store all data in the same data block. In addition, LOBs allow data to be storedin a separate segment and tablespace, or in a host file.

• LOBs can be object type attributes; LONGs cannot.

• LOBs support random piecewise access to the data through a file-like interface; LONGs arerestricted to sequential piecewise access.

The TO_LOB function can be used to covert LONG and LONG RAW values in a column to LOB values.You use this in the SELECT list of a subquery in an INSERT statement.


Contrasting LONG and LOB Data Types

LONG and LONG RAW

Single LONG column per table

Up to 2 GB

SELECT returns data

Data stored in-line

Sequential access to data

LOB

Multiple LOB columns per table

Up to 4 GB

SELECT returns locator

Data stored in-line or out-of-line

Random access to data



Anatomy of a LOB

LOB locator

The LOB column stores a locator to the LOB's value.

LOB columnof a table

LOB value

Components of a LOB

There are two distinct parts of a LOB:

• LOB value: The data that constitutes the real object being stored.

• LOB locator: A pointer to the location of the LOB value stored in the database.

Regardless of where the value of the LOB is stored, a locator is stored in the row. You can think of a LOBlocator as a pointer to the actual location of the LOB value.

A LOB column does not contain the data; it contains the locator of the LOB value.

When a user creates an internal LOB, the value is stored in the LOB segment and a locator to the out-of-line LOB value is placed in the LOB column of the corresponding row in the table. External LOBs store thedata outside the database, so only a locator to the LOB value is stored in the table.

To access and manipulate LOBs without SQL DML, you must create a LOB locator. Programmaticinterfaces operate on the LOB values, using these locators in a manner similar to operating system filehandles.

Instructor NoteAn internal LOB’s value is stored in-line with the other row data if the size of the LOB value is less than4,000 bytes. When the LOB value is larger than 4,000 bytes, the LOB value is automatically moved out ofthe row.

When you are creating a table with a LOB column, the default storage is ENABLE STORAGE IN ROW. Ifyou do not want the LOB value stored in the row, even if the size is less than 4,000 bytes, use the storageclause option DISABLE STORAGE IN ROW.



Internal LOBs

The LOB value is stored in the database.

“Four score and seven years ago

our fathers brought forth upon

this continent, a new nation,

conceived in LIBERTY, and dedicated

to the proposition that all men

are created equal.”

CLOB BLOB

Features of Internal LOBs

The internal LOB is stored inside the Oracle server. A BLOB, NCLOB, or CLOB can be one of thefollowing:

• An attribute of a user-defined type

• A column in a table

• A bind or host variable

• A PL/SQL variable, parameter, or result

Internal LOBs can take advantage of Oracle features such as:

• Concurrency mechanisms

• Redo logging and recovery mechanisms

• Transactions with commit or rollbacks

The BLOB data type is interpreted by the Oracle server as a bitstream, similar to the LONG RAW data type.

The CLOB data type is interpreted as a single-byte character stream.

The NCLOB data type is interpreted as a multiple-byte character stream, based on the byte length of thedatabase national character set.


How to Manage LOBs

Use the following method to manage an internal LOB:

1. Create and populate the table containing the LOB data type.

2. Declare and initialize the LOB locator in the program.

3. Use SELECT FOR UPDATE to lock the row containing the LOB into the LOB locator.

4. Manipulate the LOB with DBMS_LOB package procedures, OCI calls, Oracle Objects for OLE,Oracle precompilers, or JDBC using the LOB locator as a reference to the LOB value.

You can also manage LOBs through SQL.

5. Use the COMMIT command to make any changes permanent.

Instructor Note

OCI: Oracle Call Interface

Oracle Objects for OLE (OO4O): Oracle Objects for object linking and embedding

JDBC: Java Database Connectivity


Managing Internal LOBs

• To interact fully with LOB, file-like interfaces areprovided in:– PL/SQL package DBMS_LOB

– Oracle Call Interface (OCI)

– Oracle Objects for object linking and embedding(OLE)

– Pro*C/C++ and Pro*COBOL precompilers

– JDBC

• The Oracle server provides some support for LOBmanagement through SQL.



Movie(BFILE)

What Are BFILEs?

The BFILE data typesupports an external orfile-based large object as:• Attributes in an object type

• Column values in a table

What Are BFILEs?

BFILEs are external large objects (LOBs) stored in operating system files outside of the databasetablespaces. The Oracle SQL data type to support these large objects is BFILE. The BFILE data typestores a locator to the physical file. A BFILE can be in GIF, JPEG, MPEG, MPEG2, text, or otherformats. The External LOBs may be located on hard disks, CDROMs, photo CDs, or any such device,but a single LOB cannot extend from one device to another.

The BFILE data type is available so that database users can access the external file system. TheOracle9i server provides for:

• Definition of BFILE objects

• Association of BFILE objects to corresponding external files

• Security for BFILEs

The rest of the operations required to use BFILEs are possible through the DBMS_LOB package and theOracle Call Interface.

BFILEs are read-only, so they do not participate in transactions. Any support for integrity and durabilitymust be provided by the operating system. The user must create the file and place it in the appropriatedirectory, giving the Oracle process privileges to read the file. When the LOB is deleted, the Oracleserver does not delete the file. The administration of the actual files and the OS directory structures tohouse the files is the responsibility of the database administrator (DBA), system administrator, or user.The maximum size of an external large object is operating system dependent but cannot exceed fourgigabytes.

Note: BFILEs are available in the Oracle8 database and in later releases.



Securing BFILEs

User

Movie(BFILE)

Accesspermissions

Securing BFILEs

Unauthenticated access to files on a server presents a security risk. The Oracle9i Server can act as asecurity mechanism to shield the operating system from unsecured access while removing the need tomanage additional user accounts on an enterprise computer system.

File Location and Access Privileges

The file must reside on the machine where the database exists. A time-out to read a nonexistent BFILEis based on the operating system value.

You can read a BFILE in the same way as you read an internal LOB. However, there could berestrictions related to the file itself, such as:

• Access permissions

• File system space limits

• Non-Oracle manipulations of files

• OS maximum file size

The Oracle9i RDBMS does not provide transactional support on BFILEs. Any support for integrity anddurability must be provided by the underlying file system and the OS. Oracle backup and recoverymethods support only the LOB locators, not the physical BFILEs.



A New Database Object: DIRECTORY

DIRECTORY

LOB_PATH ='/oracle/lob/'

User

Movie(BFILE)

A New Database Object: DIRECTORY

A DIRECTORY is a nonschema database object that provides for administration of access and usage ofBFILEs in an Oracle9i Server.

A DIRECTORY specifies an alias for a directory on the file system of the server under which a BFILEis located. By granting suitable privileges for these items to users, you can provide secure access to filesin the corresponding directories on a user-by-user basis (certain directories can be made read-only,inaccessible, and so on).

Further, these directory aliases can be used while referring to files (open, close, read, and so on) inPL/SQL and OCI. This provides application abstraction from hard-coded path names, and givesflexibility in portably managing file locations.

The DIRECTORY object is owned by SYS and created by the DBA (or a user with CREATE ANYDIRECTORY privilege). Directory objects have object privileges, unlike any other nonschema object.Privileges to the DIRECTORY object can be granted and revoked. Logical path names are not supported.

The permissions for the actual directory are operating system dependent. They may differ from thosedefined for the DIRECTORY object and could change after the creation of the DIRECTORY object.



Guidelines for Creating DIRECTORYObjects

• Do not create DIRECTORY objects on paths withdatabase files.

• Limit the number of people who are given thefollowing system privileges:– CREATE ANY DIRECTORY

– DROP ANY DIRECTORY

• All DIRECTORY objects are owned by SYS.

• Create directory paths and properly setpermissions before using the DIRECTORY objectso that the Oracle server can read the file.

Guidelines for Creating Directory Objects

To associate an operating system file to a BFILE, you should first create a DIRECTORY object that isan alias for the full pathname to the operating system file.

Create DIRECTORY objects by using the following guidelines:

• Directories should point to paths that do not contain database files, because tampering with thesefiles could corrupt the database. Currently, only the READ privilege can be given for aDIRECTORY object.

• The system privileges CREATE ANY DIRECTORY and DROP ANY DIRECTORY should beused carefully and not granted to users indiscriminately.

• DIRECTORY objects are not schema objects; all are owned by SYS.

• Create the directory paths with appropriate permissions on the OS prior to creating theDIRECTORY object. Oracle does not create the OS path.

If you migrate the database to a different operating system, you may need to change the path value of theDIRECTORY object.

The DIRECTORY object information that you create by using the CREATE DIRECTORY command isstored in the data dictionary views DBA_DIRECTORIES and ALL_DIRECTORIES.



Managing BFILEs

• Create an OS directory and supply files.

• Create an Oracle table with a column that holdsthe BFILE data type.

• Create a DIRECTORY object.

• Grant privileges to read the DIRECTORY object tousers.

• Insert rows into the table by using the BFILENAMEfunction.

• Declare and initialize a LOB locator in a program.

• Read the BFILE.

How to Manage BFILEs

Use the following method to manage the BFILE and DIRECTORY objects:

1. Create the OS directory (as an Oracle user) and set permissions so that the Oracle server canread the contents of the OS directory. Load files into the the OS directory.

2. Create a table containing the BFILE data type in the Oracle server.

3. Create the DIRECTORY object.

4. Grant the READ privilege to it.

5. Insert rows into the table using the BFILENAME function and associate the OS files with thecorresponding row and column intersection.

6. Declare and initialize the LOB locator in a program.

7. Select the row and column containing the BFILE into the LOB locator.

8. Read the BFILE with an OCI or a DBMS_LOB function, using the locator as a reference tothe file.

Instructor NoteThere is a GRANT READ privilege so users can read from the directory created. This privilege ischecked by the DBMS_LOB package.



• Create or modify an Oracle table with a columnthat holds the BFILE data type.

• Create a DIRECTORY object by using the CREATEDIRECTORY command.

• Grant privileges to read the DIRECTORY object tousers.

ALTER TABLE employeesADD emp_video BFILE;

Preparing to Use BFILEs

CREATE DIRECTORY dir_nameAS os_path;

GRANT READ ON DIRECTORY dir_name TOuser|role|PUBLIC;

Preparing to Use BFILEs

In order to use a BFILE within an Oracle table, you need to have a table with a column of BFILE type. Forthe Oracle server to access an external file, the server needs to know the location of the file on the operatingsystem. The DIRECTORY object provides the means to specify the location of the BFILEs. Use theCREATE DIRECTORY command to specify the pointer to the location where your BFILEs are stored. Youneed the CREATE ANY DIRECTORY privilege.

Syntax Definition: CREATE DIRECTORY dir_name AS os_path;

The following commands set up a pointer to BFILEs in the system directory /$HOME/LOG_FILES andgive users the privilege to read the BFILEs from the directory.

CREATE OR REPLACE DIRECTORY log_files AS '/$HOME/LOG_FILES';GRANT READ ON DIRECTORY log_files TO PUBLIC;

In a session, the number of BFILEs that can be opened in one session is limited by the parameterSESSION_MAX_OPEN_FILES. This parameter is set in the init.ora file. Its default value is 10.

Instructor NoteTo delete a created directory, use the DROP DIRECTORY command.

Where: dir_name is the name of the directory database objectos_path is the location of the BFILEs



FUNCTION BFILENAME (directory_alias IN VARCHAR2,filename IN VARCHAR2)

RETURN BFILE;

The BFILENAME Function

Use the BFILENAME function to initialize a BFILEcolumn.

The BFILENAME Function

BFILENAME is a built-in function that initializes a BFILE column to point to an external file. Use theBFILENAME function as part of an INSERT statement to initialize a BFILE column by associating itwith a physical file in the server file system. You can use the UPDATE statement to change the referencetarget of the BFILE. A BFILE can be initialized to NULL and updated later by using the BFILENAMEfunction.

Syntax Definitions

ExampleUPDATE employees

SET emp_video = BFILENAME('LOG_FILES', 'King.avi')

WHERE employee_id = 100;

Once physical files are associated with records using SQL DML, subsequent read operations on theBFILE can be performed using the PL/SQL DBMS_LOB package and OCI. However, these files areread-only when accessed through BFILEs, and so they cannot be updated or deleted through BFILEs.

Instructor NoteYou can demonstrate this code with the code example 15_14n.sql file. This script file containsadditional statements to disable and enable triggers on EMPLOYEES table.

Where: directory_alias is the name of the DIRECTORY database object

filename is the name of the BFILE to be read



CREATE OR REPLACE PROCEDURE load_emp_bfile(p_file_loc IN VARCHAR2) IS

v_file BFILE;v_filename VARCHAR2(16);CURSOR emp_cursor ISSELECT first_name FROM employeesWHERE department_id = 60 FOR UPDATE;

BEGINFOR emp_record IN emp_cursor LOOPv_filename := emp_record.first_name || '.bmp';v_file := BFILENAME(p_file_loc, v_filename);DBMS_LOB.FILEOPEN(v_file);UPDATE employees SET emp_video = v_fileWHERE CURRENT OF emp_cursor;

DBMS_OUTPUT.PUT_LINE('LOADED FILE: '||v_filename|| ' SIZE: ' || DBMS_LOB.GETLENGTH(v_file));

DBMS_LOB.FILECLOSE(v_file);END LOOP;

END load_emp_bfile;/

Loading BFILEs

Loading BFILEsExampleLoad a BFILE pointer to an image of each employee into the EMPLOYEES table by using the DBMS_LOBpackage. The images are .bmp files stored in the /home/LOG_FILES directory.

Executing the procedure yields the following results:

EXECUTE load_emp_bfile('LOG_FILES')

Instructor Note

Although the DIRECTORY object, represented by the DIRECTORY_ALIAS parameter toBFILENAME(), need not already be defined before BFILENAME() is called by a SQL or PL/SQLprogram, the DIRECTORY object and operating system file must exist by the time you actually use theBFILE locator. For example, when the BFILE locator is used as a parameter to one of the followingoperations: OCILobFileOpen(), DBMS_LOB.FILEOPEN(), OCILOBOPEN(),DBMS_LOB.OPEN()

You can demonstrate this code with the 15_15s.sql and 15_15n.sql files. Run the 15_15s.sqlscript first.



Loading BFILEs

Use the DBMS_LOB.FILEEXISTS function to vefirythat the file exists in the operating system. The functionreturns 0 if the file does not exist, and returns 1 if thefile does exist.

CREATE OR REPLACE PROCEDURE load_emp_bfile(p_file_loc IN VARCHAR2)ISv_file BFILE; v_filename VARCHAR2(16);v_file_exists BOOLEAN;CURSOR emp_cursor IS ...

BEGINFOR emp_record IN emp_cursor LOOPv_filename := emp_record.first_name || '.bmp';v_file := BFILENAME (p_file_loc, v_filename);v_file_exists := (DBMS_LOB.FILEEXISTS(v_file) = 1);IF v_file_exists THENDBMS_LOB.FILEOPEN (v_file); ...

Using DBMS_LOB.FILEEXISTS

This function finds out whether a given BFILE locator points to a file that actually exists on the server'sfile system. This is the specification for the function:

Syntax Definitions

FUNCTION DBMS_LOB.FILEEXISTS

(file_loc IN BFILE)

RETURN INTEGER;

If the FILE_LOC parameter contains an invalid value, one of three exceptions may be raised.

In the example in the slide, the output of the DBMS_LOB.FILEEXISTS function is compared withvalue 1 and the result is returned to the BOOLEAN variable V_FILE_EXISTS.

Where: file_loc is name of the BFILE locator

RETURN INTEGER returns 0 if the physical file does not existreturns 1 if the physical file exists

Exception Nam e DescriptionNOEXIST_DIRECTORY The directory does not exist.

NOPRIV_DIRECTORY You do not have privileges for the directory.

INVALID_DIRECTORY The directory was invalidated after the file was opened.



Migrating from LONG to LOB

The Oracle9i server allows migration of LONG columns toLOB columns.

• Data migration consists of the procedure to moveexisting tables containing LONG columns to use LOBs.

• Application migration consists of changing existing LONGapplications for using LOBs.

ALTER TABLE [<schema>.] <table_name>MODIFY (<long_col_name> {CLOB | BLOB | NCLOB}

Migrating from LONG to LOB

Oracle9i Server supports the LONG-to-LOB migration using API.

Data migration: Where existing tables that contain LONG columns need to be moved to use LOB columns.This can be done using the ALTER TABLE command. In Oracle8i, an operator named TO_LOB had to beused to copy a LONG to a LOB. In Oracle9i, this operation can be performed using the syntax shown in theslide.

You can use the syntax shown to:

• Modify a LONG column to a CLOB or an NCLOB column

• Modify a LONG RAW column to a BLOB column

The constraints of the LONG column (NULL and NOT-NULL are the only allowed constraints) aremaintained for the new LOB columns. The default value specified for the LONG column is also copied tothe new LOB column.

For example, if you had a table with the following definition:

CREATE TABLE Long_tab (id NUMBER, long_col LONG);

you can change the LONG_COL column in table LONG_TAB to the CLOB data type as follows:

ALTER TABLE Long_tab MODIFY ( long_col CLOB );

For limitations on the LONG-to-LOB migration, refer to Oracle9i Application Developer’s Guide - LargeObjects.

Application Migration: Where the existing LONG applications change for using LOBs. You can use SQLand PL/SQL to access LONGs and LOBs. This API is provided for both OCI and PL/SQL.



Migrating From LONG to LOB

• Implicit conversion: LONG (LONG RAW) or aVARCHAR2(RAW) variable to a CLOB (BLOB) variable, andvice versa

• Explicit conversion:– TO_CLOB() converts LONG, VARCHAR2, and CHAR to CLOB

– TO_BLOB() converts LONG RAW and RAW to BLOB

• Function and Procedure Parameter Passing:– CLOBs and BLOBs as actual parameters

– VARCHAR2, LONG, RAW, and LONG RAW are formalparameters, and vice versa

• LOB data is acceptable in most of the SQL and PL/SQLoperators and built-in functions

Migrating from LONG to LOB (continued)

With the new LONG-to-LOB API introduced in Oracle9i, data from CLOB and BLOB columns can bereferenced by regular SQL and PL/SQL statements.

Implicit assignment and parameter passing: The LONG-to-LOB migration API supports assigning aCLOB (BLOB) variable to a LONG (LONG RAW) or a VARCHAR2(RAW) variable, and vice versa.

Explicit conversion functions: In PL/SQL, the following two new explicit conversion functions havebeen added in Oracle9i to convert other data types to CLOB and BLOB as part of LONG-to-LOBmigration:

• TO_CLOB() converts LONG, VARCHAR2, and CHAR to CLOB

• TO_BLOB() converts LONG RAW and RAW to BLOB

TO_CHAR() is enabled to convert a CLOB to a CHAR type.

Function and procedure parameter passing: This allows all the user-defined procedures and functions touse CLOBs and BLOBs as actual parameters where VARCHAR2, LONG, RAW, and LONG RAW are formalparameters, and vice versa.

Accessing in SQL and PL/SQL built-in functions and operators: A CLOB can be passed to SQL andPL/SQL VARCHAR2 built-in functions, behaving exactly like a VARCHAR2. Or the VARCHAR2variable can be passed into DBMS_LOB APIs acting like a LOB locator.

These details are discussed in detail later in this lesson.

For more information, refer to “Migrating from LONGs to LOBs” in Oracle9i Application Developer’sGuide - Large Objects (LOBs).


The DBMS_LOB Package

In releases prior to Oracle9i, you need to use the DBMS_LOB package for retrieving data from LOBs.

To create the DBMS_LOB package, the dbmslob.sql and prvtlob.plb scripts must be executed asSYS. The catproc.sql script executes the scripts. Then users can be granted appropriate privileges touse the package.

The package does not support any concurrency control mechanism for BFILE operations.

The user is responsible for locking the row containing the destination internal LOB before calling anysubprograms that involve writing to the LOB value. These DBMS_LOB routines do not implicitly lock therow containing the LOB.

Two constants are used in the specification of procedures in this package: LOBMAXSIZE andFILE_READONLY. These constants are used in the procedures and functions of DBMS_LOB; forexample, you can use them to achieve the maximum possible level of purity so that they can be used inSQL expressions.

Using the DBMS_LOB Routines

Functions and procedures in this package can be broadly classified into two types: mutators or observers.Mutators can modify LOB values, whereas observers can only read LOB values.

• Mutators: APPEND, COPY, ERASE, TRIM, WRITE, FILECLOSE, FILECLOSEALL, andFILEOPEN

• Observers: COMPARE, FILEGETNAME, INSTR, GETLENGTH, READ, SUBSTR, FILEEXISTS,and FILEISOPEN



• Working with LOB often requires the use of theOracle-supplied package DBMS_LOB.

• DBMS_LOB provides routines to access andmanipulate internal and external LOBs.

• Oracle9i enables retrieving LOB data directly usingSQL, without using any special LOB API.

• In PL/SQL you can define a VARCHAR2 for a CLOBand a RAW for BLOB.



• Modify LOB values:

APPEND, COPY, ERASE, TRIM, WRITE, LOADFROMFILE

• Read or examine LOB values:

GETLENGTH, INSTR, READ, SUBSTR

• Specific to BFILEs:

FILECLOSE, FILECLOSEALL, FILEEXISTS,FILEGETNAME, FILEISOPEN, FILEOPEN


The DBMS_LOB Package (continued)

APPEND Append the contents of the source LOB to the destination LOB

COPY Copy all or part of the source LOB to the destination LOB

ERASE Erase all or part of a LOB

LOADFROMFILE Load BFILE data into an internal LOB

TRIM Trim the LOB value to a specified shorter length

WRITE Write data to the LOB from a specified offset

GETLENGTH Get the length of the LOB value

INSTR Return the matching position of the nth occurrence of the pattern in the LOB

READ Read data from the LOB starting at the specified offset

SUBSTR Return part of the LOB value starting at the specified offset

FILECLOSE Close the file

FILECLOSEALL Close all previously opened files

FILEEXISTS Check if the file exists on the server

FILEGETNAME Get the directory alias and file name

FILEISOPEN Check if the file was opened using the input BFILE locators

FILEOPEN Open a file


Using the DBMS_LOB Routines

All functions in the DBMS_LOB package return NULL if any input parameters are NULL . All mutatorprocedures in the DBMS_LOB package raise an exception if the destination LOB /BFILE is input as NULL.

Only positive, absolute offsets are allowed. They represent the number of bytes or characters from thebeginning of LOB data from which to start the operation. Negative offsets and ranges observed in SQLstring functions and operators are not allowed. Corresponding exceptions are raised upon violation. Thedefault value for an offset is 1, which indicates the first byte or character in the LOB value.

Similarly, only natural number values are allowed for the amount (BUFSIZ) parameter. Negative valuesare not allowed.

Instructor NoteThe offset indicates the starting point when reading from or writing to a large object. If the LOB is a CLOBor NCLOB, this is the nth character; if it is a BLOB or BFILE, this is the nth byte.



• NULL parameters get NULL returns.

• Offsets:– BLOB, BFILE: Measured in bytes

– CLOB, NCLOB: Measured in characters

• There are no negative values for parameters.


DBMS_LOB.READ

Call the READ procedure to read and return piecewise a specified AMOUNT of data from a given LOB,starting from OFFSET. An exception is raised when no more data remains to be read from the sourceLOB. The value returned in AMOUNT will be less than the one specified, if the end of the LOB is reachedbefore the specified number of bytes or characters could be read. In the case of CLOBs, the character set ofdata in BUFFER is the same as that in the LOB.

PL/SQL allows a maximum length of 32767 for RAW and VARCHAR2 parameters. Make sure theallocated system resources are adequate to support these buffer sizes for the given number of usersessions. Otherwise, the Oracle server raises the appropriate memory exceptions.Note: BLOB and BFILE return RAW; the others return VARCHAR2.

DBMS_LOB.WRITE

Call the WRITE procedure to write piecewise a specified AMOUNT of data into a given LOB, from theuser-specified BUFFER, starting from an absolute OFFSET from the beginning of the LOB value.

Make sure (especially with multibyte characters) that the amount in bytes corresponds to the amount ofbuffer data. WRITE has no means of checking whether they match, and will write AMOUNT bytes of thebuffer contents into the LOB.

Instructor NoteThese two procedures are given as examples of the routines used from DBMS_LOB. Depending on theaudience, you can go through more of the routines using ALL_SOURCE:

SELECT text FROM ALL_SOURCE WHERE name = 'DBMS_LOB';


DBMS_LOB.READ and DBMS_LOB.WRITE

PROCEDURE READ (

lobsrc IN BFILE|BLOB|CLOB ,

amount IN OUT BINARY_INTEGER,

offset IN INTEGER,

buffer OUT RAW|VARCHAR2 )

PROCEDURE WRITE (

lobdst IN OUT BLOB|CLOB,

amount IN OUT BINARY_INTEGER,

offset IN INTEGER := 1,

buffer IN RAW|VARCHAR2 ) -- RAW for BLOB



Adding LOB Columnsto a Table

ALTER TABLE employees ADD(resume CLOB,picture BLOB);

Adding LOB Columns to a Table

LOB columns are defined by way of SQL data definition language (DDL), as in the ALTER TABLEstatement in the slide. The contents of a LOB column is stored in the LOB segment, whereas the column inthe table contains only a reference to that specific storage area, called the LOB locator. In PL/SQL you candefine a variable of type LOB, which contains only the value of the LOB locator.


Populating LOB Columns

You can insert a value directly into a LOB column by using host variables in SQL or in PL/SQL,3GL-embedded SQL, or OCI.

You can use the special functions EMPTY_BLOB and EMPTY_CLOB in INSERT or UPDATE statementsof SQL DML to initialize a NULL or non-NULL internal LOB to empty. These are available as specialfunctions in Oracle SQL DML, and are not part of the DBMS_LOB package.

Before you can start writing data to an internal LOB using OCI or the DBMS_LOB package, the LOBcolumn must be made nonnull, that is, it must contain a locator that points to an empty or populated LOBvalue. You can initialize a BLOB column's value to empty by using the function EMPTY_BLOB in theVALUES clause of an INSERT statement. Similarly, a CLOB or NCLOB column's value can be initializedby using the function EMPTY_CLOB.

The result of using the function EMPTY_CLOB() or EMPTY_BLOB() means that the LOB is initialized,but not populated with data. To populate the LOB column, you can use an update statement.

You can use an INSERT statement to insert a new row and populate the LOB column at the same time.

When you create a LOB instance, the Oracle server creates and places a locator to the out-of-line LOBvalue in the LOB column of a particular row in the table. SQL, OCI, and other programmatic interfacesoperate on LOBs through these locators.


Insert a row into a table with LOB columns:

Initialize a LOB column using the EMPTY_BLOB() function:

Populating LOB Columns

INSERT INTO employees (employee_id, first_name,last_name, email, hire_date, job_id,salary, resume, picture)

VALUES (405, 'Marvin', 'Ellis', 'MELLIS', SYSDATE,'AD_ASST', 4000, EMPTY_CLOB(),NULL);

UPDATE employeesSET resume = 'Date of Birth: 8 February 1951',

picture = EMPTY_BLOB()WHERE employee_id = 405;


Populating LOB Columns (continued)

The EMPTY_B/CLOB() function can be used as a DEFAULT column constraint, as in the examplebelow. This initializes the LOB columns with locators.

CREATE TABLE emp_hiredata

(employee_id NUMBER(6),

first_name VARCHAR2(20),

last_name VARCHAR2(25),resume CLOB DEFAULT EMPTY_CLOB(),

picture BLOB DEFAULT EMPTY_BLOB());



UPDATE CLOB column

Updating LOB by Using SQL

UPDATE employeesSET resume = 'Date of Birth: 1 June 1956'WHERE employee_id = 170;

Updating LOB by Using SQL

You can update a LOB column by setting it to another LOB value, to NULL, or by using the emptyfunction appropriate for the LOB data type (EMPTY_CLOB() or EMPTY_BLOB()). You can updatethe LOB using a bind variable in embedded SQL, the value of which may be NULL, empty, orpopulated. When you set one LOB equal to another, a new copy of the LOB value is created. Theseactions do not require a SELECT FOR UPDATE statement. You must lock the row prior to the updateonly when updating a piece of the LOB.


Updating LOB by Using DBMS_LOB in PL/SQL

In the example in the slide, the LOBLOC variable serves as the LOB locator, and the AMOUNT variable isset to the length of the text you want to add. The SELECT FOR UPDATE statement locks the row andreturns the LOB locator for the RESUME LOB column. Finally, the PL/SQL package procedure WRITE iscalled to write the text into the LOB value at the specified offset. WRITEAPPEND appends to the existingLOB value.

The example shows how to fetch a CLOB column in releases before Oracle9i. In those releases, it was notpossible to fetch a CLOB column directly into a character column. The column value needed to be boundto a LOB locator, which is accessed by the DBMS_LOB package. An example later in this lesson showsthat you can directly fetch a CLOB column by binding it to a character variable.

Note: In versions prior to Oracle9i, Oracle did not allow LOBs in the WHERE clause of UPDATE andSELECT. Now SQL functions of LOBs are allowed in predicates of WHERE. An example is shown later inthis lesson.

Instructor NoteEMPTY_B/CLOB() is not the same as NULL, and an IS NULL test on a LOB initialized EMPTY returnsFALSE.


Updating LOB by Using DBMS_LOB inPL/SQL

DECLARElobloc CLOB; -- serves as the LOB locatortext VARCHAR2(32767):='Resigned: 5 August 2000';amount NUMBER ; -- amount to be writtenoffset INTEGER; -- where to start writing

BEGINSELECT resume INTO loblocFROM employeesWHERE employee_id = 405 FOR UPDATE;offset := DBMS_LOB.GETLENGTH(lobloc) + 2;amount := length(text);DBMS_LOB.WRITE (lobloc, amount, offset, text );text := ' Resigned: 30 September 2000';SELECT resume INTO loblocFROM employeesWHERE employee_id = 170 FOR UPDATE;amount := length(text);DBMS_LOB.WRITEAPPEND(lobloc, amount, text);COMMIT;

END;



Selecting CLOB Values by Using SQL

SELECT employee_id, last_name , resume -- CLOBFROM employeesWHERE employee_id IN (405, 170);

Selecting CLOB Values by Using SQL

It is possible to see the data in a CLOB column by using a SELECT statement. It is not possible tosee the data in a BLOB or BFILE column by using a SELECT statement in iSQL*Plus. You have touse a tool that can display binary information for a BLOB, as well as the relevant software for aBFILE; for example, you can use Oracle Forms.

Instructor NoteUse the iSQL*Plus command SET LONG n to display more data, if the information in the LOBcolumn is truncated.



Selecting CLOB Values by Using DBMS_LOB

• DBMS_LOB.SUBSTR(lob_column, no_of_chars, starting)

• DBMS_LOB.INSTR (lob_column, pattern)

SELECT DBMS_LOB.SUBSTR (resume, 5, 18),DBMS_LOB.INSTR (resume,' = ')

FROM employeesWHERE employee_id IN (170, 405);

Selecting CLOB Values by Using DBMS_LOB

DBMS_LOB.SUBSTR

Use DBMS_LOB.SUBSTR to display part of a LOB. It is similar in functionality to the SQL functionSUBSTR.

DBMS_LOB.INSTR

Use DBMS_LOB.INSTR to search for information within the LOB. This function returns thenumerical position of the information.

Note: Starting with Oracle9i, you can also use SQL functions SUBSTR and INSTR to perform theoperations shown in the slide.

Instructor NoteThis works in iSQL*Plus because the LOB column is a CLOB and not a BLOB or BFILE.

You can use DBMS_LOB.INSTR to do searches for information in the WHERE clause of a SELECTstatement. You may not want do this for performance reasons.

Unlike a LONG, where you can only select data and it is not possible to do any form of manipulationor searching, DBMS_LOB.INSTR/SUBSTR highlights a distinct difference between usage of LONGsand that of LOBs.



Selecting CLOB Values in PL/SQL

DECLAREtext VARCHAR2(4001);

BEGINSELECT resume INTO textFROM employeesWHERE employee_id = 170;DBMS_OUTPUT.PUT_LINE('text is: '|| text);END;/

Selecting CLOB Values in PL/SQL

The slide shows the code for accessing CLOB values that can be implicitly converted to VARCHAR2 inOracle9i. The value of the column RESUME, when selected into a VARCHAR2 variable TEXT, isimplicitly converted.

In prior releases, to access a CLOB column, first you must to retrieve the CLOB column value into aCLOB variable and specify the amount and offset size. Then you use the DBMS_LOB package to read theselected value. The code using DBMS_LOB is as follows:DECLARE

rlob clob;text VARCHAR2(4001);amt number := 4001;offset number := 1;

BEGINSELECT resume INTO rlobFROM employeesWHERE employee_id = 170;DBMS_LOB.READ(rlob, amt, offset, text);DBMS_OUTPUT.PUT_LINE('text is: '|| text);

END;/


Removing LOBs

A LOB instance can be deleted (destroyed) using appropriate SQL DML statements. The SQL statementDELETE deletes a row and its associated internal LOB value. To preserve the row and destroy only thereference to the LOB, you must update the row, by replacing the LOB column value with NULL or anempty string, or by using the EMPTY_B/CLOB() function.

Note: Replacing a column value with NULL and using EMPTY_B/CLOB are not the same. Using NULLsets the value to null, using EMPTY_B/CLOB ensures there is nothing in the column.

A LOB is destroyed when the row containing the LOB column is deleted when the table is dropped ortruncated, or implicitly when all the LOB data is updated.

You must explicitly remove the file associated with a BFILE using operating system commands.

To erase part of an internal LOB, you can use DBMS_LOB.ERASE.


Delete a row containing LOBs:

Disassociate a LOB value from a row:

Removing LOBs

DELETEFROM employeesWHERE employee_id = 405;

UPDATE employeesSET resume = EMPTY_CLOB()WHERE employee_id = 170;



Temporary LOBs

• Temporary LOBs:– Provide an interface to support creation of LOBs

that act like local variables

– Can be BLOBs, CLOBs, or NCLOBs

– Are not associated with a specific table

– Are created using DBMS_LOB.CREATETEMPORARYprocedure

– Use DBMS_LOB routines

• The lifetime of a temporary LOB is a session.

• Temporary LOBs are useful for transforming datain permanent internal LOBs.

Temporary LOBs

Temporary LOBs provide an interface to support the creation and deletion of LOBs that act like localvariables. Temporary LOBs can be BLOBs, CLOBs, or NCLOBs.

Features of temporary LOBs:

• Data is stored in your temporary tablespace, not in tables.

• Temporary LOBs are faster than persistent LOBs because they do not generate any redo or rollbackinformation.

• Temporary LOBs lookup is localized to each user’s own session; only the user who creates atemporary LOB can access it, and all temporary LOBs are deleted at the end of the session inwhich they were created.

• You can create a temporary LOB using DBMS_LOB.CREATETEMPORARY.

Temporary LOBs are useful when you want to perform some transformational operation on a LOB, forexample, changing an image type from GIF to JPEG. A temporary LOB is empty when created and doesnot support the EMPTY_B/CLOB functions.

Use the DBMS_LOB package to use and manipulate temporary LOBs.

Instructor NoteTemporary LOBs are introduced in Oracle8i.

The default lifetime of a temporary LOB is a session. It can also be cleaned up at the end of the call; thisis specified at the time of creation.



PL/SQL procedure to create and test a temporary LOB:

Creating a Temporary LOB

CREATE OR REPLACE PROCEDURE IsTempLOBOpen(p_lob_loc IN OUT BLOB, p_retval OUT INTEGER)

ISBEGIN-- create a temporary LOBDBMS_LOB.CREATETEMPORARY (p_lob_loc, TRUE);-- see if the LOB is open: returns 1 if openp_retval := DBMS_LOB.ISOPEN (p_lob_loc);DBMS_OUTPUT.PUT_LINE ('The file returned a value

....' || p_retval);-- free the temporary LOBDBMS_LOB.FREETEMPORARY (p_lob_loc);

END;/

Creating a Temporary LOB

The example in the slide shows a user-defined PL/SQL procedure, IsTempLOBOpen, that creates atemporary LOB. This procedure accepts a LOB locator as input, creates a temporary LOB, opens it, andtests whether the LOB is open.

The IsTempLOBOpen procedure uses the procedures and functions from the DBMS_LOB package asfollows:

• The CREATETEMPORARY procedure is used to create the temporary LOB.

• The ISOPEN function is used to test whether a LOB is open: this function returns the value 1 ifthe LOB is open.

• The FREETEMPORARY procedure is used to free the temporary LOB; memory increasesincrementally as the number of temporary LOBs grows, and you can reuse temporary LOB spacein your session by explicitly freeing temporary LOBs.



Summary


• Identify four built-in types for large objects: BLOB,CLOB, NCLOB, and BFILE

• Describe how LOBs replace LONG and LONG RAW

• Describe two storage options for LOBs:– The Oracle server (internal LOBs)

– External host files (external LOBs)

• Use the DBMS_LOB PL/SQL package to provideroutines for LOB management

• Use temporary LOBs in a session

SummaryThere are four LOB data types:

• A BLOB is a binary large object.

• A CLOB is a character large object.

• A NCLOB stores multibyte national character set data.

• A BFILE is a large object stored in a binary file outside the database.

LOBs can be stored internally (in the database) or externally (in an operating system file). You canmanage LOBs by using the DBMS_LOB package and its procedures.

Temporary LOBs provide an interface to support the creation and deletion of LOBs that act like localvariables.





• Creating object types, using the new data typesCLOB and BLOB

• Creating a table with LOB data types as columns

• Using the DBMS_LOB package to populate andinteract with the LOB data

Practice 15 OverviewIn this practice you create a table with both BLOB and CLOB columns. Then, you use the DBMS_LOBpackage to populate the table and manipulate the data.


Practice 151. Create a table called PERSONNEL by executing the script file lab15_1.sql. The table contains

the following attributes and data types:

2. Insert two rows into the PERSONNEL table, one each for employees 2034 and 2035. Use the emptyfunction for the CLOB, and provide NULL as the value for the BLOB.

3. Examine and execute the script lab15_3.sql. The script creates a table named REVIEW_TABLE.This table contains annual review information for each employee. The script also contains twostatements to insert review details for two employees.

4. Update the PERSONNEL table.

a. Populate the CLOB for the first row, using the following subquery in a SQL UPDATE statement:SELECT ann_reviewFROM review_table


b. Populate the CLOB for the second row, using PL/SQL and the DBMS_LOB package. Use thefollowing SELECT statement to provide a value.

SELECT ann_review

FROM review_table


Column Name Data type Length

ID NUMBER 6

last_name VARCHAR2 35

review CLOB N/A

picture BLOB N/A


Practice 15 (continued)

If you have time

5. Create a procedure that adds a locator to a binary file into the PICTURE column of the COUNTRIEStable. The binary file is a picture of the country. The image files are named after the country IDs. Youneed to load an image file locator into all rows in Europe region (REGION_ID = 1) in theCOUNTRIES table. The DIRECTORY object name that stores a pointer to the location of the binaryfiles is called COUNTRY_PIC. This object is already created for you.

a. Use the command below to add the image column to the COUNTRIES table (or uselab15_5_add.sql)

ALTER TABLE countries ADD (picture BFILE);

b. Create a PL/SQL procedure called load_country_image that reads a locator intoyour picture column. Have the program test to see if the file exists, using the functionDBMS_LOB.FILEEXISTS. If the file is not existing, your procedure should display amessage that the file can not be opened. Have your program report information about theload to the screen.

c. Invoke the procedure by passing the name of the directory object COUNTRY_PIC asparameter. Note that you should pass the directory object in single quotation marks.

Sample output follows:

Instructor NoteThe class needs to add a BFILE column to the COUNTRIES table to complete question 5 of Practice 15.The script is stored in lab\lab15_5_add.sql.

This practice question 5 requires the instructor to run the file lab\lab15_5_setup.sql. The scriptcreates the DIRECTORY objects and grants READ permissions to these objects. Edit the file and fill in thedatabase connection for the classroom. Also, fill in the file location on the server to identify where theCOUNTRY_PIC and LOG_FILES are located.

The steps in this script must be followed before the students can complete question 5 of Practice 15. Afterrunning the above script, you can verify that the directory alias is created by querying theALL_DIRECTORIES data dictionary view as follows:

SELECT directory_name, directory_path FROM all_directories;


60 minutes Practice

120 minutes Total


Creating Database Triggers



Objectives


• Describe different types of triggers

• Describe database triggers and their use

• Create database triggers

• Describe database trigger firing rules

• Remove database triggers

Lesson Aim

In this lesson, you learn how to create and use database triggers.



Types of Triggers

A trigger:

• Is a PL/SQL block or a PL/SQL procedureassociated with a table, view, schema, or thedatabase

• Executes implicitly whenever a particular eventtakes place

• Can be either:– Application trigger: Fires whenever an event occurs

with a particular application

– Database trigger: Fires whenever a data event (suchas DML) or system event (such as logon orshutdown) occurs on a schema or database

Types of Triggers

Application triggers execute implicitly whenever a particular data manipulation language (DML)event occurs within an application. An example of an application that uses triggers extensively is onedeveloped with Oracle Forms Developer.

Database triggers execute implicitly when a data event such as DML on a table (an INSERT,UPDATE, or DELETE triggering statement), an INSTEAD OF trigger on a view, or data definitionlanguage (DDL) statements such as CREATE and ALTER are issued, no matter which user isconnected or which application is used. Database triggers also execute implicitly when some useractions or database system actions occur, for example, when a user logs on, or the DBA shut downsthe database.

Note: Database triggers can be defined on tables and on views. If a DML operation is issued on aview, the INSTEAD OF trigger defines what actions take place. If these actions include DMLoperations on tables, then any triggers on the base tables are fired.

Database triggers can be system triggers on a database or a schema. With a database, triggers fire foreach event for all users; with a schema, triggers fire for each event for that specific user.

This course covers creating database triggers. Creating database triggers based on system events isdiscussed in the lesson “More Trigger Concepts.”

Instructor Note

Database administration type triggers have been introduced in Oracle8i. Some of the triggers that canbe created include user actions of logging on and off, and database shutdown and startup.

These types of triggers are covered in a later lesson.



Guidelines for Designing Triggers

• Design triggers to:– Perform related actions

– Centralize global operations

• Do not design triggers:– Where functionality is already built into the Oracle

server

– That duplicate other triggers

• Create stored procedures and invoke them in atrigger, if the PL/SQL code is very lengthy.

• The excessive use of triggers can result incomplex interdependencies, which may be difficultto maintain in large applications.

Guidelines for Designing Triggers• Use triggers to guarantee that when a specific operation is performed, related actions are

performed.

• Use database triggers only for centralized, global operations that should be fired for the triggeringstatement, regardless of which user or application issues the statement.

• Do not define triggers to duplicate or replace the functionality already built into the Oracledatabase. For example do not define triggers to implement integrity rules that can be done by usingdeclarative constraints. An easy way to remember the design order for a business rule is to:

– Use built-in constraints in the Oracle server such as, primary key, foreign key and so on

– Develop a database trigger or develop an application such as a servlet or EnterpriseJavaBean (EJB) on your middle tier

– Use a presentation interface such as Oracle Forms, dynamic HTML, Java ServerPages(JSP) and so on, if you cannot develop your business rule as mentioned above, which mightbe a presentation rule.

• The excessive use of triggers can result in complex interdependencies, which may be difficult tomaintain in large applications. Only use triggers when necessary, and beware of recursive andcascading effects.

• If the logic for the trigger is very lengthy, create stored procedures with the logic and invoke themin the trigger body.

• Note that database triggers fire for every user each time the event occurs on which the trigger iscreated.



Application

INSERT INTO EMPLOYEES. . .;

EMPLOYEES table CHECK_SAL trigger

Database Trigger: Example

…

Example of a Database TriggerIn this example, the database trigger CHECK_SAL checks salary values whenever any application triesto insert a row into the EMPLOYEES table. Values that are out of range according to the job categorycan be rejected, or can be allowed and recorded in an audit table.


• Recursive trigger: This is a trigger that contains a DML operation changing the very same table.

• Cascading trigger: The action of one trigger cascades to another trigger, causing this secondtrigger to fire. The Oracle server allows up to 32 triggers to cascade at any one time. However,the number of cascading triggers can be limited by changing the value of the OPEN_CURSORSdatabase initialization parameter, which is set to 50 by default.



Creating DML Triggers

A triggering statement contains:

• Trigger timing– For table: BEFORE, AFTER

– For view: INSTEAD OF

• Triggering event: INSERT, UPDATE, or DELETE

• Table name: On table, view

• Trigger type: Row or statement

• WHEN clause: Restricting condition

• Trigger body: PL/SQL block

Database Trigger

Before coding the trigger body, decide on the values of the components of the trigger: the triggertiming, the triggering event, and the trigger type.

If multiple triggers are defined for a table, be aware that the order in which multiple triggers of thesame type fire is arbitrary. To ensure that triggers of the same type are fired in a particular order,consolidate the triggers into one trigger that calls separate procedures in the desired order.

Part Description Possible Values

Trigger timing When the trigger fires in relation to thetriggering event

BEFOREAFTERINSTEAD OF

Triggering event Which data manipulation operation on thetable or view causes the trigger to fire

INSERTUPDATEDELETE

Trigger type How many times the trigger bodyexecutes

StatementRow

Trigger body What action the trigger performs Complete PL/SQL block



DML Trigger Components

Trigger timing: When should the trigger fire?• BEFORE: Execute the trigger body before the

triggering DML event on a table.

• AFTER: Execute the trigger body after thetriggering DML event on a table.

• INSTEAD OF: Execute the trigger body instead ofthe triggering statement. This is used for viewsthat are not otherwise modifiable.

BEFORE Triggers

This type of trigger is frequently used in the following situations:

• To determine whether that triggering statement should be allowed to complete. (This situationenables you to eliminate unnecessary processing of the triggering statement and its eventualrollback in cases where an exception is raised in the triggering action.)

• To derive column values before completing a triggering INSERT or UPDATE statement.

• To initialize global variables or flags, and to validate complex business rules.

AFTER TriggersThis type of trigger is frequently used in the following situations:

• To complete the triggering statement before executing the triggering action.• To perform different actions on the same triggering statement if a BEFORE trigger is already

present.

INSTEAD OF TriggersThis type of trigger is used to provide a transparent way of modifying views that cannot be modifieddirectly through SQL DML statements because the view is not inherently modifiable.You can write INSERT, UPDATE, and DELETE statements against the view. The INSTEAD OFtrigger works invisibly in the background performing the action coded in the trigger body directly onthe underlying tables.




Triggering user event: Which DML statement causesthe trigger to execute? You can use any of thefollowing:• INSERT

• UPDATE

• DELETE

The Triggering EventThe triggering event or statement can be an INSERT, UPDATE, or DELETE statement on a table.

• When the triggering event is an UPDATE statement, you can include a column list to identifywhich columns must be changed to fire the trigger. You cannot specify a column list for anINSERT or for a DELETE statement, because they always affect entire rows.

. . . UPDATE OF salary . . .

• The triggering event can contain one, two, or all three of these DML operations.

. . . INSERT or UPDATE or DELETE

. . . INSERT or UPDATE OF job_id . . .




Trigger type: Should the trigger body execute for eachrow the statement affects or only once?

• Statement: The trigger body executes once for thetriggering event. This is the default. A statementtrigger fires once, even if no rows are affected at all.

• Row: The trigger body executes once for each rowaffected by the triggering event. A row trigger is notexecuted if the triggering event affects no rows.

Statement Triggers and Row Triggers

You can specify that the trigger will be executed once for every row affected by the triggeringstatement (such as a multiple row UPDATE) or once for the triggering statement, no matter how manyrows it affects.

Statement Trigger

A statement trigger is fired once on behalf of the triggering event, even if no rows are affected at all.

Statement triggers are useful if the trigger action does not depend on the data from rows that areaffected or on data provided by the triggering event itself: for example, a trigger that performs acomplex security check on the current user.

Row Trigger

A row trigger fires each time the table is affected by the triggering event. If the triggering event affectsno rows, a row trigger is not executed.

Row triggers are useful if the trigger action depends on data of rows that are affected or on dataprovided by the triggering event itself.




Trigger body: What action should the trigger perform?

The trigger body is a PL/SQL block or a call to aprocedure.

Trigger Body

The trigger action defines what needs to be done when the triggering event is issued. The PL/SQL blockcan contain SQL and PL/SQL statements, and can define PL/SQL constructs such as variables, cursors,exceptions, and so on. You can also call a PL/SQL procedure or a Java procedure.

Additionally, row triggers use correlation names to access the old and new column values of the rowbeing processed by the trigger.

Note: The size of a trigger cannot be more than 32 K.

Instructor Note

The Java procedure needs to be encapsulated in a PL/SQL wrapper. To call a Java procedure, you use theCALL statement. The syntax for the CALL statement to call a PL/SQL procedure is shown in a laterlesson.

Before creating any triggers, run the catproc.sql script while connected as SYS. This scriptautomatically runs all of the scripts required for, or used within, the procedural extensions to the Oracleserver.



Firing Sequence

Triggering actionBEFORE statementtrigger

BEFORE row triggerAFTER row trigger

AFTER statement trigger

DML statement

Use the following firing sequence for a trigger on atable, when a single row is manipulated:

INSERT INTO departments (department_id,department_name, location_id)

VALUES (400, 'CONSULTING', 2400);

…

Creating Row or Statement Triggers

Create a statement trigger or a row trigger based on the requirement that the trigger must fire once foreach row affected by the triggering statement, or just once for the triggering statement, regardless ofthe number of rows affected.

When the triggering data manipulation statement affects a single row, both the statement trigger andthe row trigger fire exactly once.

Example

This SQL statement does not differentiate statement triggers from row triggers, because exactly onerow is inserted into the table using this syntax.



UPDATE employeesSET salary = salary * 1.1WHERE department_id = 30;

Firing Sequence

Use the following firing sequence for a trigger on atable, when many rows are manipulated:

BEFORE statement trigger

BEFORE row triggerAFTER row trigger...

BEFORE row triggerAFTER row trigger...

AFTER statement trigger

Creating Row or Statement Triggers (continued)

When the triggering data manipulation statement affects many rows, the statement trigger fires exactlyonce, and the row trigger fires once for every row affected by the statement.

Example

The SQL statement in the slide above causes a row-level trigger to fire a number of times equal to thenumber of rows that satisfy the WHERE clause, that is, the number of employees reporting todepartment 30.



Syntax for CreatingDML Statement Triggers

CREATE [OR REPLACE] TRIGGER trigger_nametiming

event1 [OR event2 OR event3]ON table_name

trigger_body

Note: Trigger names must be unique with respect toother triggers in the same schema.

Syntax:

Syntax for Creating a Statement Trigger

Trigger names must be unique with respect to other triggers in the same schema. Trigger names do notneed to be unique with respect to other schema objects, such as tables, views, and procedures.

Using column names along with the UPDATE clause in the trigger improves performance, because thetrigger fires only when that particular column is updated and thus avoids unintended firing when anyother column is updated.

trigger name Is the name of the trigger

timing Indicates the time when the trigger fires in relation to thetriggering event:BEFOREAFTER

event Identifies the data manipulation operation that causes thetrigger to fire:INSERTUPDATE [OF column]DELETE

table/view_name Indicates the table associated with the trigger

trigger body Is the trigger body that defines the action performed by thetrigger, beginning with either DECLARE or BEGIN, endingwith END, or a call to a procedure



Creating DML Statement Triggers

CREATE OR REPLACE TRIGGER secure_empBEFORE INSERT ON employeesBEGINIF (TO_CHAR(SYSDATE,'DY') IN ('SAT','SUN')) OR

(TO_CHAR(SYSDATE,'HH24:MI')NOT BETWEEN '08:00' AND '18:00')

THEN RAISE_APPLICATION_ERROR (-20500,'You mayinsert into EMPLOYEES table only

during business hours.');END IF;

END;/

Example:

Creating DML Statement TriggersYou can create a BEFORE statement trigger in order to prevent the triggering operation fromsucceeding if a certain condition is violated.

For example, create a trigger to restrict inserts into the EMPLOYEES table to certain business hours,Monday through Friday.

If a user attempts to insert a row into the EMPLOYEES table on Saturday, the user sees the message,the trigger fails, and the triggering statement is rolled back. Remember that theRAISE_APPLICATION_ERROR is a server-side built-in procedure that returns an error to the userand causes the PL/SQL block to fail.

When a database trigger fails, the triggering statement is automatically rolled back by the Oracleserver.

Instructor Note

The time specified in the trigger in the slide is based on the database server system time. Hence, if youare in a different time zone from the database server, your DML may not work even if your systemclock is within the time specified in the code.



Testing SECURE_EMP

INSERT INTO employees (employee_id, last_name,first_name, email, hire_date,job_id, salary, department_id)

VALUES (300, 'Smith', 'Rob', 'RSMITH', SYSDATE,'IT_PROG', 4500, 60);

ExampleInsert a row into the EMPLOYEES table during nonbusiness hours. When the date and time are out ofthe business timings specified in the trigger, you get the error message as shown in the slide.

Instructor Note

Note that the row might be inserted if you are in a different timezone from the database server. Thetrigger fires even if your system clock is within these business hours.



Using Conditional Predicates

CREATE OR REPLACE TRIGGER secure_empBEFORE INSERT OR UPDATE OR DELETE ON employeesBEGINIF (TO_CHAR (SYSDATE,'DY') IN ('SAT','SUN')) OR

(TO_CHAR (SYSDATE, 'HH24') NOT BETWEEN '08' AND '18')THEN

IF DELETING THENRAISE_APPLICATION_ERROR (-20502,'You may delete from

EMPLOYEES table only during business hours.');ELSIF INSERTING THEN

RAISE_APPLICATION_ERROR (-20500,'You may insert intoEMPLOYEES table only during business hours.');

ELSIF UPDATING ('SALARY') THENRAISE_APPLICATION_ERROR (-20503,'You may update

SALARY only during business hours.');ELSE

RAISE_APPLICATION_ERROR (-20504,'You may updateEMPLOYEES table only during normal hours.');

END IF;END IF;

END;

Combining Triggering Events

You can combine several triggering events into one by taking advantage of the special conditionalpredicates INSERTING, UPDATING, and DELETING within the trigger body.

Example

Create one trigger to restrict all data manipulation events on the EMPLOYEES table to certain businesshours, Monday through Friday.



Creating a DML Row Trigger


event1 [OR event2 OR event3]ON table_name

[REFERENCING OLD AS old | NEW AS new]FOR EACH ROW

[WHEN (condition)]trigger_body

Syntax:

Syntax for Creating a Row Trigger

trigger_name Is the name of the trigger

timing Indicates the time when the trigger fires in relation to the triggering event:BEFOREAFTERINSTEAD OF

event Identifies the data manipulation operation that causes the trigger to fire:INSERTUPDATE [OF column]DELETE

table_name Indicates the table associated with the triggerREFERENCING Specifies correlation names for the old and new values of the current row

(The default values are OLD and NEW)

FOR EACH ROW Designates that the trigger is a row trigger

WHEN Specifies the trigger restriction; (This conditional predicate must beenclosed in parenthesis and is evaluated for each row to determine whetheror not the trigger body is executed.)

trigger body Is the trigger body that defines the action performed by the trigger,beginning with either DECLARE or BEGIN, ending with END, or a call to aprocedure



Creating DML Row Triggers

CREATE OR REPLACE TRIGGER restrict_salaryBEFORE INSERT OR UPDATE OF salary ON employeesFOR EACH ROWBEGINIF NOT (:NEW.job_id IN ('AD_PRES', 'AD_VP'))

AND :NEW.salary > 15000THENRAISE_APPLICATION_ERROR (-20202,'Employee

cannot earn this amount');END IF;

END;/

Creating a Row TriggerYou can create a BEFORE row trigger in order to prevent the triggering operation from succeeding if acertain condition is violated.

Create a trigger to allow only certain employees to be able to earn a salary of more than 15,000.

If a user attempts to do this, the trigger raises an error.

UPDATE employees

SET salary = 15500

WHERE last_name = 'Russell';

Instructor Note

Alternatively, the value 15,000 in the program could be a value in a table that can be retrieved into theprogram, or a global variable.



Using OLD and NEW Qualifiers

CREATE OR REPLACE TRIGGER audit_emp_values

AFTER DELETE OR INSERT OR UPDATE ON employees

FOR EACH ROW

BEGIN

INSERT INTO audit_emp_table (user_name, timestamp,

id, old_last_name, new_last_name, old_title,

new_title, old_salary, new_salary)

VALUES (USER, SYSDATE, :OLD.employee_id,

:OLD.last_name, :NEW.last_name, :OLD.job_id,

:NEW.job_id, :OLD.salary, :NEW.salary );

END;

/

Using OLD and NEW Qualifiers

Within a ROW trigger, reference the value of a column before and after the data change by prefixing itwith the OLD and NEW qualifier.

• The OLD and NEW qualifiers are available only in ROW triggers.

• Prefix these qualifiers with a colon (:) in every SQL and PL/SQL statement.

• There is no colon (:) prefix if the qualifiers are referenced in the WHEN restricting condition.

Note: Row triggers can decrease the performance if you do a lot of updates on larger tables.

Instructor Note

If you run the above code and get a compilation error, check whether you have a table calledAUDIT_EMP_TABLE; if not, run 16_addtabs.sql to create this table. Point out to the studentsthat the OLD and NEW are host variable created implicitly. The same rules apply to these host variablesin PL/SQL bodies as in anonymous PL/SQL blocks.

Data Operation Old Value New Value

INSERT NULL Inserted value

UPDATE Value before update Value after update

DELETE Value before delete NULL



Using OLD and NEW Qualifiers:Example Using Audit_Emp_Table

INSERT INTO employees(employee_id, last_name, job_id, salary, ...)

VALUES (999, 'Temp emp', 'SA_REP', 1000, ...);

UPDATE employees

SET salary = 2000, last_name = 'Smith'


SELECT user_name, timestamp, ... FROM audit_emp_table

Using OLD and NEW Qualifiers: Example Using AUDIT_EMP_TABLE

Create a trigger on the EMPLOYEES table to add rows to a user table, AUDIT_EMP_TABLE, logginga user’s activity against the EMPLOYEES table. The trigger records the values of several columns bothbefore and after the data changes by using the OLD and NEW qualifiers with the respective columnname.

There is additional column COMMENTS in the AUDIT_EMP_TABLE that is not shown in this slide.



Restricting a Row Trigger

CREATE OR REPLACE TRIGGER derive_commission_pctBEFORE INSERT OR UPDATE OF salary ON employeesFOR EACH ROWWHEN (NEW.job_id = 'SA_REP')

BEGINIF INSERTING

THEN :NEW.commission_pct := 0;ELSIF :OLD.commission_pct IS NULL

THEN :NEW.commission_pct := 0;ELSE

:NEW.commission_pct := :OLD.commission_pct + 0.05;END IF;

END;/

ExampleTo restrict the trigger action to those rows that satisfy a certain condition, provide a WHEN clause.

Create a trigger on the EMPLOYEES table to calculate an employee’s commission when a row isadded to the EMPLOYEES table, or when an employee’s salary is modified.

The NEW qualifier cannot be prefixed with a colon in the WHEN clause because the WHEN clause isoutside the PL/SQL blocks.

Instructor NoteTo assign values to columns using the NEW qualifier, create BEFORE ROW triggers. If you attempt tocreate an AFTER ROW trigger for the above code, you get a compilation error:CREATE OR REPLACE TRIGGER derive_commission_pctAFTER INSERT OR UPDATE OF salary ON employeesFOR EACH ROW WHEN (NEW.job_id = 'SA_REP')BEGIN

IF INSERTING THEN :NEW.commission_pct := 0;ELSIF :OLD.commission_pct IS NULL

THEN :NEW.commission_pct := 0;ELSE :NEW.commission_pct :=

:OLD.commission_pct * (:NEW.salary/:OLD.salary);END IF;

END;/

CREATE OR REPLACE TRIGGER derive_commission_pct*ERROR at line 1:ORA-04084: cannot change NEW values for this trigger type



INSTEAD OF Triggers

Application

INSERT INTO my_view. . .;

MY_VIEW

INSTEAD OFTrigger

INSERTTABLE1

UPDATETABLE2

INSTEAD OF Triggers

Use INSTEAD OF triggers to modify data in which the DML statement has been issued against aninherently nonupdatable view. These triggers are called INSTEAD OF triggers because, unlike othertriggers, the Oracle server fires the trigger instead of executing the triggering statement. This trigger isused to perform an INSERT, UPDATE, or DELETE operation directly on the underlying tables.

You can write INSERT, UPDATE, or DELETE statements against a view, and the INSTEAD OFtrigger works invisibly in the background to make the right actions take place.

Why Use INSTEAD OF Triggers?

A view cannot be modified by normal DML statements if the view query contains set operators, groupfunctions, clauses such as GROUP BY, CONNECT BY, START, the DISTINCT operator, or joins.For example, if a view consists of more than one table, an insert to the view may entail an insertioninto one table and an update to another. So, you write an INSTEAD OF trigger that fires when youwrite an insert against the view. Instead of the original insertion, the trigger body executes, whichresults in an insertion of data into one table and an update to another table.

Note: If a view is inherently updateable and has INSTEAD OF triggers, the triggers take precedence.INSTEAD OF triggers are row triggers.

The CHECK option for views is not enforced when insertions or updates to the view are performed byusing INSTEAD OF triggers. The INSTEAD OF trigger body must enforce the check.

Instructor Note

You can update a subset of join views.



Creating an INSTEAD OF Trigger

CREATE [OR REPLACE] TRIGGER trigger_nameINSTEAD OFevent1 [OR event2 OR event3]ON view_name

[REFERENCING OLD AS old | NEW AS new][FOR EACH ROW]trigger_body

Syntax:

Syntax for Creating an INSTEAD OF Trigger

Note: INSTEAD OF triggers can be written only for views. BEFORE and AFTER options are notvalid.

trigger_name Is the name of the trigger.

INSTEAD OF Indicates that the trigger belongs to a view

event Identifies the data manipulation operation that causes the triggerto fire:INSERTUPDATE [OF column]DELETE

view_name Indicates the view associated with trigger

REFERENCING Specifies correlation names for the old and new values of thecurrent row (The defaults are OLD and NEW)

FOR EACHROW

Designates the trigger to be a row trigger; INSTEAD OF triggerscan only be row triggers: if this is omitted, the trigger is stilldefined as a row trigger.

trigger body Is the trigger body that defines the action performed by thetrigger, beginning with either DECLARE or BEGIN, and endingwith END or a call to a procedure



Example:

The following example creates two new tables, NEW_EMPS and NEW_DEPTS, based on theEMPLOYEES and DEPARTMENTS tables respectively. It also creates an EMP_DETAILS view fromthe EMPLOYEES and DEPARTMENTS tables. The example also creates an INSTEAD OF trigger,NEW_EMP_DEPT. When a row is inserted into the EMP_DETAILS view, instead of inserting the rowdirectly into the view, rows are added into the NEW_EMPS and NEW_DEPTS tables, based on the datain the INSERT statement. Similarly, when a row is modified or deleted through the EMP_DETAILSview, corresponding rows in the NEW_EMPS and NEW_DEPTS tables are affected.CREATE TABLE new_emps AS

SELECT employee_id, last_name, salary, department_id,email, job_id, hire_date

FROM employees;

CREATE TABLE new_depts ASSELECT d.department_id, d.department_name, d.location_id,

sum(e.salary) tot_dept_salFROM employees e, departments d

WHERE e.department_id = d.department_id

GROUP BY d.department_id, d.department_name, d.location_id;

CREATE VIEW emp_details ASSELECT e.employee_id, e.last_name, e.salary, e.department_id,

e.email, e.job_id, d.department_name, d.location_id

FROM employees e, departments d

WHERE e.department_id = d.department_id;

CREATE OR REPLACE TRIGGER new_emp_dept

INSTEAD OF INSERT OR UPDATE OR DELETE ON emp_details

FOR EACH ROW

BEGIN

IF INSERTING THEN

INSERT INTO new_emps

VALUES (:NEW.employee_id, :NEW.last_name, :NEW.salary,:NEW.department_id, :NEW.email, :New.job_id, SYSDATE);

UPDATE new_depts

SET tot_dept_sal = tot_dept_sal + :NEW.salary

WHERE department_id = :NEW.department_id;

ELSIF DELETING THEN

DELETE FROM new_emps

WHERE employee_id = :OLD.employee_id;

UPDATE new_depts

SET tot_dept_sal = tot_dept_sal - :OLD.salary

WHERE department_id = :OLD.department_id;


Creating an INSTEAD OF Trigger (continued)

Example:

ELSIF UPDATING ('salary')THEN

UPDATE new_emps

SET salary = :NEW.salary


UPDATE new_depts

SET tot_dept_sal = tot_dept_sal + (:NEW.salary - :OLD.salary)


ELSIF UPDATING ('department_id')

THEN

UPDATE new_emps

SET department_id = :NEW.department_id


UPDATE new_depts

SET tot_dept_sal = tot_dept_sal - :OLD.salary


UPDATE new_deptsSET tot_dept_sal = tot_dept_sal + :NEW.salary

WHERE department_id = :NEW.department_id;

END IF;

END;

/

Note: This example is explained in the next page by using graphics.

Instructor NoteDemonstration: 16_inst_of.sql

Purpose: Creates the two new tables, a view and the trigger. The script shows the data in the view andthe tables. A direct INSERT into the view implicitly executes the trigger, which INSERTs into theNEW_EMPS table and UPDATEs the NEW_DEPTS table.

The trigger is not complete. It is programmed to handle the INSERT and UPDATE of the SALARY orDEPARTMENT_ID, and a DELETE. Any other form of UPDATE is not programmed.




INSERT INTO emp_details(employee_id, ... )VALUES(9001,'ABBOTT',3000,10,'abbott.mail.com','HR_MAN');

INSERT into EMP_DETAILS that is based on EMPLOYEES andDEPARTMENTS tables

INSTEAD OF INSERTinto EMP_DETAILS

…

1


You can create an INSTEAD OF trigger in order to maintain the base tables on which a view is based.

Assume that an employee name will be inserted using the view EMP_DETAILS that is created basedon the EMPLOYEES and DEPARTMENTS tables. Create a trigger that results in the appropriateINSERT and UPDATE to the base tables. The slide in the next page explains how an INSTEAD OFTRIGGER behaves in this situation.




INSERT INTO emp_details(employee_id, ... )VALUES(9001,'ABBOTT',3000,10,'abbott.mail.com','HR_MAN');

INSERT into EMP_DETAILS that is based on EMPLOYEES andDEPARTMENTS tables

INSTEAD OF INSERTinto EMP_DETAILS

INSERT intoNEW_EMPS

UPDATENEW_DEPTS

…

……

1

2 3


Because of the INSTEAD OF TRIGGER on the view EMP_DETAILS, instead of inserting the newemployee record into the EMPLOYEES table:

• A row is inserted into the NEW_EMPS table.

• The TOTAL_DEPT_SAL column of the NEW_DEPTS table is updated. The salary valuesupplied for the new employee is added to the existing total salary of the department to whichthe new employee has been assigned.



Differentiating Between Database Triggersand Stored Procedures

Triggers

Defined with CREATE TRIGGER

Data dictionary contains sourcecode in USER_TRIGGERS

Implicitly invoked

COMMIT, SAVEPOINT, andROLLBACK are not allowed

Procedures

Defined with CREATE PROCEDURE

Data dictionary contains source codein USER_SOURCE

Explicitly invoked

COMMIT, SAVEPOINT, and ROLLBACK

are allowed

Database Triggers and Stored Procedures

There are differences between database triggers and stored procedures:

Triggers are fully compiled when the CREATE TRIGGER command is issued and the P code isstored in the data dictionary.

If errors occur during the compilation of a trigger, the trigger is still created.

Database Trigger Stored Procedure

Invoked implicitly Invoked explicitly

COMMIT, ROLLBACK, andSAVEPOINT statements are not allowedwithin the trigger body. It is possible tocommit or rollback indirectly by calling aprocedure, but it is not recommendedbecause of side effects to transactions.

COMMIT, ROLLBACK, and SAVEPOINTstatements are permitted within the procedurebody.



Differentiating Between Database Triggersand Form Builder Triggers

INSERT INTO EMPLOYEES. . .;

EMPLOYEES table CHECK_SAL trigger

BEFOREINSERTrow

…

Differences between a Database Trigger and a Form Builder Trigger

Database triggers are different from Form Builder triggers.

Database Trigger Form Builder Trigger

Executed by actions from any database toolor application

Executed only within a particular Form Builderapplication

Always triggered by a SQL DML, DDL, or acertain database action

Can be triggered by navigating from field to field, bypressing a key, or by many other actions

Is distinguished as either a statement or rowtrigger

Is distinguished as a statement or row trigger

Upon failure, causes the triggering statementto roll back

Upon failure, causes the cursor to freeze and maycause the entire transaction to roll back

Fires independently of, and in addition to,Form Builder triggers

Fires independently of, and in addition to, databasetriggers

Executes under the security domain of theauthor of the trigger

Executes under the security domain of the FormBuilder user



ALTER TRIGGER trigger_name DISABLE | ENABLE

Managing Triggers

Disable or reenable a database trigger:

ALTER TABLE table_name DISABLE | ENABLE ALL TRIGGERS

Disable or reenable all triggers for a table:

ALTER TRIGGER trigger_name COMPILE

Recompile a trigger for a table:

Trigger Modes: Enabled or Disabled

• When a trigger is first created, it is enabled automatically.

• The Oracle server checks integrity constraints for enabled triggers and guarantees that triggerscannot compromise them. In addition, the Oracle server provides read-consistent views forqueries and constraints, manages the dependencies, and provides a two-phase commit process ifa trigger updates remote tables in a distributed database.

• Disable a specific trigger by using the ALTER TRIGGER syntax, or disable all triggers on atable by using the ALTER TABLE syntax.

• Disable a trigger to improve performance or to avoid data integrity checks when loadingmassive amounts of data by using utilities such as SQL*Loader. You may also want to disablethe trigger when it references a database object that is currently unavailable, owing to a failednetwork connection, disk crash, offline data file, or offline tablespace.

Compile a Trigger• Use the ALTER TRIGGER command to explicitly recompile a trigger that is invalid.

• When you issue an ALTER TRIGGER statement with the COMPILE option, the triggerrecompiles, regardless of whether it is valid or invalid.



DROP TRIGGER Syntax

To remove a trigger from the database, use the DROPTRIGGER syntax:

DROP TRIGGER trigger_name;

DROP TRIGGER secure_emp;

Example:

Note: All triggers on a table are dropped when thetable is dropped.

Removing Triggers

When a trigger is no longer required, you can use a SQL statement in iSQL*Plus to drop it.



Trigger Test Cases

• Test each triggering data operation, as well asnontriggering data operations.

• Test each case of the WHEN clause.

• Cause the trigger to fire directly from a basic dataoperation, as well as indirectly from a procedure.

• Test the effect of the trigger upon other triggers.

• Test the effect of other triggers upon the trigger.

Testing Triggers

• Ensure that the trigger works properly by testing a number of cases separately.

• Take advantage of the DBMS_OUTPUT procedures to debug triggers. You can also use theProcedure Builder debugging tool to debug triggers. Using Procedure Builder is discussed inAppendix F, “Creating Program Units by Using Procedure Builder.”

Instructor Note

When writing new application code with procedures and functions, you should verify theUSER_TRIGGERS data dictionary view for any existing triggers. The newly added procedures andfunctions may fire these triggers implicitly, which is not intended. Hence, test the procedures andfunctions, and drop any triggers that are not required.



Trigger Execution Modeland Constraint Checking

1. Execute all BEFORE STATEMENT triggers.

2. Loop for each row affected:a. Execute all BEFORE ROW triggers.

b. Execute all AFTER ROW triggers.

3. Execute the DML statement and perform integrityconstraint checking.

4. Execute all AFTER STATEMENT triggers.

Trigger Execution ModelA single DML statement can potentially fire up to four types of triggers: BEFORE and AFTERstatement and row triggers. A triggering event or a statement within the trigger can cause one or moreintegrity constraints to be checked. Triggers can also cause other triggers to fire (cascading triggers).

All actions and checks done as a result of a SQL statement must succeed. If an exception is raisedwithin a trigger and the exception is not explicitly handled, all actions performed because of theoriginal SQL statement are rolled back. This includes actions performed by firing triggers. Thisguarantees that integrity constraints can never be compromised by triggers.

When a trigger fires, the tables referenced in the trigger action may undergo changes by other users'transactions. In all cases, a read-consistent image is guaranteed for modified values the trigger needsto read (query) or write (update).

Instructor Note

The example in the following slide demonstrates the constraint verification procedure. You candemonstrate this concept by using the file 16_trig_constr.sql.



Trigger Execution Model and ConstraintChecking: Example

CREATE OR REPLACE TRIGGER constr_emp_trigAFTER UPDATE ON employeesFOR EACH ROW

BEGININSERT INTO departments

VALUES (999, 'dept999', 140, 2400);END;/

UPDATE employees SET department_id = 999


-- Successful after trigger is fired

UPDATE employees SET department_id = 999

WHERE employee_id = 170;-- Integrity constraint violation error

Trigger Execution Model and Constraint Checking: Example

The example in the slide explains a situation in which the integrity constraint can be taken care of byusing a trigger. Table EMPLOYEES has a foreign key constraint on the DEPARTMENT_ID column ofthe DEPARTMENTS table.

In the first SQL statement, the DEPARTMENT_ID of the employee with EMPLOYEE_ID 170 ismodified to 999.

Because such a department does not exist in the DEPARTMENTS table, the statement raises theexception -2292 for the integrity constraint violation.

A trigger CONSTR_EMP_TRIG is created that inserts a new department 999 into the DEPARTMENTStable.

When the UPDATE statement that modifies the department of employee 170 to 999 is issued, thetrigger fires. Then, the foreign key constraint is checked. Because the trigger inserted the department999 into the DEPARTMENTS table, the foreign key constraint check is successful and there is noexception.

This process works with Oracle8i and later releases. The example described in the slide produces arun-time error in releases prior to Oracle8i.



VAR_PACKpackage

AUDIT_EMP_TRIGFOR EACH ROWIncrement variables

2

A Sample Demonstration for TriggersUsing Package Constructs

AUDIT_EMP_TABAFTER STATEMENTCopy and then resetvariables

3

DML intoEMPLOYEES table

1

AUDIT_TABLE4

A Sample Demonstration

The following pages of PL/SQL subprograms are an example of the interaction of triggers, packagedprocedures, functions, and global variables.

The sequence of events:

1. Issue an INSERT, UPDATE, or DELETE command that can manipulate one or many rows.

2. AUDIT_EMP_TRIG, the AFTER ROW trigger, calls the packaged procedure to increment theglobal variables in the package VAR_PACK. Because this is a row trigger, the trigger fires oncefor each row that you updated.

3. When the statement has finished, AUDIT_EMP_TAB, the AFTER STATEMENT trigger, calls theprocedure AUDIT_EMP.

4. This procedure assigns the values of the global variables into local variables using the packagedfunctions, updates the AUDIT_TABLE, and then resets the global variables.

Instructor Note

The slide presents a simplified version of the demonstration. First run the script16_drop_trg.sql. The following demonstration scripts contain a complete explanation about thedemonstration as described in the next few pages.

Demonstration: 16_comp_ex_spec.sql and 16_comp_ex_body.sql

Purpose: To create the VAR_PACK package, the AUDIT_EMP procedure, and theAUDIT_EMP_TRIG trigger. If you wish to complement the code by running an example, insert a rowinto AUDIT_TABLE with your user name and EMPLOYEES as the table name, and then write anINSERT, UPDATE, or DELETE command to see the triggers working.



After Row and After Statement Triggers

CREATE OR REPLACE TRIGGER audit_emp_trigAFTER UPDATE or INSERT or DELETE on EMPLOYEESFOR EACH ROWBEGINIF DELETING THEN var_pack.set_g_del(1);ELSIF INSERTING THEN var_pack.set_g_ins(1);ELSIF UPDATING ('SALARY')

THEN var_pack.set_g_up_sal(1);ELSE var_pack.set_g_upd(1);END IF;

END audit_emp_trig;/

CREATE OR REPLACE TRIGGER audit_emp_tabAFTER UPDATE or INSERT or DELETE on employeesBEGINaudit_emp;

END audit_emp_tab;/

AFTER Row and AFTER Statement Triggers

The trigger AUDIT_EMP_TRIG is a row trigger that fires after every row manipulated. This triggerinvokes the package procedures depending on the type of DML performed. For example, if the DMLupdates salary of an employee, then the trigger invokes the procedure SET_G_UP_SAL. This packageprocedure inturn invokes the function G_UP_SAL. This function increments the package variableGV_UP_SAL that keeps account of the number of rows being changed due to update of the salary.

The trigger AUDIT_EMP_TAB will fire after the statement has finished. This trigger invokes theprocedure AUDIT_EMP, which is on the following pages. The AUDIT_EMP procedure updates theAUDIT_TABLE table. An entry is made into the AUDIT_TABLE table with the information such asthe user who performed the DML, the table on which DML is performed, and the total number of suchdata manipulations performed so far on the table (indicated by the value of the corresponding columnin the AUDIT_TABLE table). At the end, the AUDIT_EMP procedure resets the package variables to0.

Instructor Note

This is not the only way to implement this business requirement, but it does show the way packages,triggers, procedures, and packaged variables can interact.



Demonstration: VAR_PACK PackageSpecification

CREATE OR REPLACE PACKAGE var_packIS-- these functions are used to return the-- values of package variablesFUNCTION g_del RETURN NUMBER;FUNCTION g_ins RETURN NUMBER;FUNCTION g_upd RETURN NUMBER;FUNCTION g_up_sal RETURN NUMBER;

-- these procedures are used to modify the-- values of the package variablesPROCEDURE set_g_del (p_val IN NUMBER);PROCEDURE set_g_ins (p_val IN NUMBER);PROCEDURE set_g_upd (p_val IN NUMBER);PROCEDURE set_g_up_sal (p_val IN NUMBER);

END var_pack;/

var_pack.sql

Demonstration: VAR_PACK Package Bodyvar_pack_body.sqlCREATE OR REPLACE PACKAGE BODY var_pack IS

gv_del NUMBER := 0; gv_ins NUMBER := 0;gv_upd NUMBER := 0; gv_up_sal NUMBER := 0;

FUNCTION g_del RETURN NUMBER ISBEGIN

RETURN gv_del;END;

FUNCTION g_ins RETURN NUMBER ISBEGIN

RETURN gv_ins;END;

FUNCTION g_upd RETURN NUMBER ISBEGIN

RETURN gv_upd;END;

FUNCTION g_up_sal RETURN NUMBER ISBEGIN

RETURN gv_up_sal;END;

(continued on the next page)


VAR_PACK Package Body (continued)

PROCEDURE set_g_del (p_val IN NUMBER) ISBEGIN

IF p_val = 0 THENgv_del := p_val;

ELSE gv_del := gv_del +1;END IF;

END set_g_del;PROCEDURE set_g_ins (p_val IN NUMBER) ISBEGIN

IF p_val = 0 THENgv_ins := p_val;

ELSE gv_ins := gv_ins +1;END IF;

END set_g_ins;PROCEDURE set_g_upd (p_val IN NUMBER) ISBEGIN

IF p_val = 0 THENgv_upd := p_val;

ELSE gv_upd := gv_upd +1;END IF;

END set_g_upd;PROCEDURE set_g_up_sal (p_val IN NUMBER) ISBEGIN

IF p_val = 0 THENgv_up_sal := p_val;

ELSE gv_up_sal := gv_up_sal +1;END IF;

END set_g_up_sal;END var_pack;/



CREATE OR REPLACE PROCEDURE audit_emp ISv_del NUMBER := var_pack.g_del;v_ins NUMBER := var_pack.g_ins;v_upd NUMBER := var_pack.g_upd;v_up_sal NUMBER := var_pack.g_up_sal;

BEGINIF v_del + v_ins + v_upd != 0 THEN

UPDATE audit_table SETdel = del + v_del, ins = ins + v_ins,upd = upd + v_upd

WHERE user_name=USER AND tablename='EMPLOYEES'AND column_name IS NULL;

END IF;IF v_up_sal != 0 THEN

UPDATE audit_table SET upd = upd + v_up_salWHERE user_name=USER AND tablename='EMPLOYEES'AND column_name = 'SALARY';

END IF;-- resetting global variables in package VAR_PACK

var_pack.set_g_del (0); var_pack.set_g_ins (0);var_pack.set_g_upd (0); var_pack.set_g_up_sal (0);

END audit_emp;

Demonstration: Using theAUDIT_EMP Procedure

Updating the AUDIT_TABLE with the AUDIT_EMP Procedure

The AUDIT_EMP procedure updates the AUDIT_TABLE and calls the functions in the packageVAR_PACK that reset the package variables, ready for the next DML statement.



Procedure Package Trigger

xxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxx



Summary


Localvariable

Develop different types of procedural database constructs depending on their usage.

Construct Usage

Procedure PL/SQL programming block that is stored in the database for repeatedexecution

Package Group of related procedures, functions, variables, cursors, constants, andexceptions

Trigger PL/SQL programming block that is executed implicitly by a data manipulationstatement





• Creating statement and row triggers

• Creating advanced triggers to add to thecapabilities of the Oracle database


You create statement and row triggers in this practice. You create procedures that will be invokedfrom the triggers.


Practice 16

1. Changes to data are allowed on tables only during normal office hours of 8:45 a.m. until 5:30p.m., Monday through Friday.

Create a stored procedure called SECURE_DML that prevents the DML statement fromexecuting outside of normal office hours, returning the message, “You may only makechanges during normal office hours.”

2. a. Create a statement trigger on the JOBS table that calls the above procedure.

b. Test the procedure by temporarily modifying the hours in the procedure and attempting toinsert a new record into the JOBS table. (Example: replace 08:45 with 16:45; This attemptresults in an error message)

After testing, reset the procedure hours asspecified in question 1 and recreate theprocedure as in question 1 above.

If you have time:

3. Employees should receive an automatic increase in salary if the minimum salary for a job isincreased. Implement this requirement through a trigger on the JOBS table.a. Create a stored procedure named UPD_EMP_SAL to update the salary amount. This

procedure accepts two parameters: the job ID for which salary has to be updated, and thenew minimum salary for this job ID. This procedure is executed from the trigger on theJOBS table.

b. Create a row trigger named UPDATE_EMP_SALARY on the JOBS table that invokes theprocedure UPD_EMP_SAL, when the minimum salary in the JOBS table is updated for aspecified job ID.

c. Query the EMPLOYEES table to see the current salary for employees who areprogrammers.

d. Increase the minimum salary for the Programmer job from 4,000 to 5,000.

e. Employee Lorentz (employee ID 107) had a salary of less than 4,500. Verify that hersalary has been increased to the new minimum of 5,000.

Instructor Note

If you encounter errors while performing practice question 3d, verify that you do not have thetriggers trig_constr_emp and constr_emp_trig in your schema. These triggers arecreated and dropped when you run the demo 16_trig_constr.sql and code example16_34s.sql. Students did not create these triggers earlier.


45 minutes Practice

95 minutes Total


More Trigger Concepts



Objectives


• Create additional database triggers

• Explain the rules governing triggers

• Implement triggers

Lesson Aim

In this lesson, you learn how to create more database triggers and learn the rules governing triggers.You also learn many applications of triggers.

Instructor Note

This lesson contains several sections with discrete concepts that do not tie in with one another.

Towards the end of this lesson, there are many practical examples of triggers. You do not need to gothrough each example. Gauge the level of detail to the audience requirements.




• Triggering user event:– CREATE, ALTER, or DROP

– Logging on or off

• Triggering database or system event:– Shutting down or starting up the database

– A specific error (or any error) being raised


Before coding the trigger body, decide on the components of the trigger.

Triggers on system events can be defined at the database or schema level. For example, a databaseshutdown trigger is defined at the database level. Triggers on data definition language (DDL)statements, or a user logging on or off, can also be defined at either the database level or schema level.Triggers on DML statements are defined on a specific table or a view.

A trigger defined at the database level fires for all users, and a trigger defined at the schema or tablelevel fires only when the triggering event involves that schema or table.

Triggering events that can cause a trigger to fire:

• A data definition statement on an object in the database or schema

• A specific user (or any user) logging on or off

• A database shutdown or startup

• A specific or any error that occurs

Instructor Note

The trigger concepts described above were introduced in Oracle8i and are intended primarily for use byDBAs.



Creating Triggers on DDL Statements


[ddl_event1 [OR ddl_event2 OR ...]]ON {DATABASE|SCHEMA}

trigger_body

Syntax:

Create Trigger Syntax

The trigger body represents a complete PL/SQL block.

You can create triggers for these events on DATABASE or SCHEMA. You also specify BEFORE orAFTER for the timing of the trigger.

DDL triggers fire only if the object being created is a cluster, function, index, package, procedure, role,sequence, synonym, table, tablespace, trigger, type, view, or user.

Instructor Note

The Oracle server fires the trigger in the existing user transaction.

DDL_Event Possible Values

CREATE Causes the Oracle server to fire the trigger whenever a CREATE statementadds a new database object to the dictionary

ALTER Causes the Oracle server to fire the trigger whenever an ALTER statementmodifies a database object in the data dictionary

DROP Causes the Oracle server to fire the trigger whenever a DROP statementremoves a database object in the data dictionary



Creating Triggers on System Events


[database_event1 [OR database_event2 OR ...]]ON {DATABASE|SCHEMA}

trigger_body

Create Trigger Syntax

You can create triggers for these events on DATABASE or SCHEMA except SHUTDOWN and STARTUP,which apply only to the DATABASE.

Instructor Note

For each of these triggering events, the Oracle server opens an autonomous transaction scope, fires thetrigger, and commits any separate transaction (regardless of any existing user transaction).

An autonomous transaction scope is an independent transaction that can be committed without affectingthe existing transaction. This was introduced in Oracle8i. Autonomous transactions are not covered inthis course.

Database_event Possible Values

AFTERSERVERERROR

Causes the Oracle server to fire the trigger whenever a server error message islogged

AFTER LOGON Causes the Oracle server to fire the trigger whenever a user logs on to thedatabase

BEFORE LOGOFF Causes the Oracle server to fire the trigger whenever a user logs off the database

AFTER STARTUP Causes the Oracle server to fire the trigger whenever the database is opened

BEFORESHUTDOWN

Causes the Oracle server to fire the trigger whenever the database is shut down



LOGON and LOGOFF Trigger Example

CREATE OR REPLACE TRIGGER logon_trigAFTER LOGON ON SCHEMABEGININSERT INTO log_trig_table(user_id, log_date, action)VALUES (USER, SYSDATE, 'Logging on');END;/

CREATE OR REPLACE TRIGGER logoff_trigBEFORE LOGOFF ON SCHEMABEGININSERT INTO log_trig_table(user_id, log_date, action)VALUES (USER, SYSDATE, 'Logging off');END;/

LOGON and LOGOFF Trigger Example

You can create this trigger to monitor how often you log on and off, or you may want to write a reportthat monitors the length of time for which you are logged on. When you specify ON SCHEMA, thetrigger fires for the specific user. If you specify ON DATABASE, the trigger fires for all users.

Instructor NoteSample LOG_TRIG_TABLE:

DESCRIBE log_trig_table

SELECT * FROM log_trig_table;



CALL Statements

CREATE OR REPLACE TRIGGER log_employeeBEFORE INSERT ON EMPLOYEES

CALL log_execution/

CREATE [OR REPLACE] TRIGGER trigger_nametimingevent1 [OR event2 OR event3]ON table_name

[REFERENCING OLD AS old | NEW AS new][FOR EACH ROW][WHEN condition]

CALL procedure_name

CALL Statements

A CALL statement enables you to call a stored procedure, rather than coding the PL/SQL body in thetrigger itself. The procedure can be implemented in PL/SQL, C, or Java.

The call can reference the trigger attributes :NEW and :OLD as parameters as in the following example:

CREATE TRIGGER salary_check

BEFORE UPDATE OF salary, job_id ON employees

FOR EACH ROW

WHEN (NEW.job_id <> 'AD_PRES')

CALL check_sal(:NEW.job_id, :NEW.salary)

/

Note: There is no semicolon at the end of the CALL statement.

In the example above, the trigger calls a procedure check_sal. The procedure compares the newsalary with the salary range for the new job ID from the JOBS table.

Instructor NoteCode example 17_7n.sql creates a procedure check_sal that compares the new salary with therange of the salary allowed for that job ID. The script also contains the above code for the trigger andstatements to fire the trigger. The trigger is disabled at the end of the script of the next page.

If you receive compilation errors while creating the trigger shown in the slide, re-create theLOG_EXECTION procedure by running the script demo\09_logexec.sql.



Trigger event

UPDATE employeesSET salary = 3400WHERE last_name = 'Stiles';

EMPLOYEES tableFailure

Triggered table/mutating table

BEFORE UPDATE row

CHECK_SALARYtrigger

Reading Datafrom a Mutating Table

…… 3400

Rules Governing Triggers

Reading and writing data using triggers is subject to certain rules. The restrictions apply only to rowtriggers, unless a statement trigger is fired as a result of ON DELETE CASCADE.

Mutating TableA mutating table is a table that is currently being modified by an UPDATE, DELETE, or INSERTstatement, or a table that might need to be updated by the effects of a declarative DELETE CASCADEreferential integrity action. A table is not considered mutating for STATEMENT triggers.

The triggered table itself is a mutating table, as well as any table referencing it with the FOREIGN KEYconstraint. This restriction prevents a row trigger from seeing an inconsistent set of data.

Instructor Note

Before Oracle8i, it was not possible to change data in the primary key, foreign key, or unique keycolumns of a constraining table. This rule no longer applies in Oracle8i.



Mutating Table: Example

CREATE OR REPLACE TRIGGER check_salaryBEFORE INSERT OR UPDATE OF salary, job_idON employeesFOR EACH ROWWHEN (NEW.job_id <> 'AD_PRES')

DECLAREv_minsalary employees.salary%TYPE;v_maxsalary employees.salary%TYPE;

BEGINSELECT MIN(salary), MAX(salary)INTO v_minsalary, v_maxsalaryFROM employeesWHERE job_id = :NEW.job_id;IF :NEW.salary < v_minsalary OR

:NEW.salary > v_maxsalary THENRAISE_APPLICATION_ERROR(-20505,'Out of range');

END IF;END;/

Mutating Table: ExampleThe CHECK_SALARY trigger in the example, attempts to guarantee that whenever a new employee isadded to the EMPLOYEES table or whenever an existing employee’s salary or job ID is changed, theemployee’s salary falls within the established salary range for the employee’s job.

When an employee record is updated, the CHECK_SALARY trigger is fired for each row that is updated.The trigger code queries the same table that is being updated. Hence, it is said that the EMPLOYEEStable is mutating table.



Mutating Table: Example

UPDATE employees

SET salary = 3400

WHERE last_name = 'Stiles';

Mutating Table: Example (continued)

Try to read from a mutating table.

If you restrict the salary within a range between the minimum existing value and the maximum existingvalue you get a run-time error. The EMPLOYEES table is mutating, or in a state of change; therefore,the trigger cannot read from it.

Remember that functions can also cause a mutating table error when they are invoked in a DMLstatement.

Instructor NoteA mutating table is a table that is currently being modified by an UPDATE, DELETE, or INSERTstatement. A mutating table cannot be changed, because the resulting DML could change data that is inan inconsistent state.



Implementing Triggers

You can use trigger for:

• Security

• Auditing

• Data integrity

• Referential integrity

• Table replication

• Computing derived data automatically

• Event logging

Implementing Triggers

Develop database triggers in order to enhance features that cannot otherwise be implemented by theOracle server or as alternatives to those provided by the Oracle server.

Feature Enhancement

Security The Oracle server allows table access to users or roles. Triggers allowtable access according to data values.

Auditing The Oracle server tracks data operations on tables. Triggers trackvalues for data operations on tables.

Data integrity The Oracle server enforces integrity constraints. Triggers implementcomplex integrity rules.

Referential integrity The Oracle server enforces standard referential integrity rules. Triggersimplement nonstandard functionality.

Table replication The Oracle server copies tables asynchronously into snapshots.Triggers copy tables synchronously into replicas.

Derived data The Oracle server computes derived data values manually. Triggerscompute derived data values automatically.

Event logging The Oracle server logs events explicitly. Triggers log eventstransparently.



GRANT SELECT, INSERT, UPDATE, DELETEON employeesTO clerk; -- database roleGRANT clerk TO scott;

Controlling Security Withinthe Server

Controlling Security Within the Server

Develop schemas and roles within the Oracle server to control the security of data operations on tablesaccording to the identity of the user.

• Base privileges upon the username supplied when the user connects to the database.

• Determine access to tables, views, synonyms, and sequences.

• Determine query, data manipulation, and data definition privileges.

Instructor NoteThere is no code example script file for the example in the slide.



CREATE OR REPLACE TRIGGER secure_empBEFORE INSERT OR UPDATE OR DELETE ON employees

DECLAREv_dummy VARCHAR2(1);

BEGINIF (TO_CHAR (SYSDATE, 'DY') IN ('SAT','SUN'))THEN RAISE_APPLICATION_ERROR (-20506,'You may only

change data during normal business hours.');END IF;SELECT COUNT(*) INTO v_dummy FROM holidayWHERE holiday_date = TRUNC (SYSDATE);IF v_dummy > 0 THEN RAISE_APPLICATION_ERROR(-20507,

'You may not change data on a holiday.');END IF;END;/

Controlling Securitywith a Database Trigger

Controlling Security With a Database Trigger

Develop triggers to handle more complex security requirements.

• Base privileges on any database values, such as the time of day, the day of the week, and so on.

• Determine access to tables only.

• Determine data manipulation privileges only.

Instructor Note

The code example script file for the example in the slide contains additional statements that create theHOLIDAY table and insert values into the table.



AUDIT INSERT, UPDATE, DELETEON departmentsBY ACCESS

WHENEVER SUCCESSFUL;

Using the Server Facility toAudit Data Operations

The Oracle server stores the audit information in adata dictionary table or operating system file.

Auditing Data OperationsYou can audit data operations within the Oracle server. Database auditing is used to monitor and gatherdata about specific database activities. The DBA can gather statistics about which tables are beingupdated, how many I/Os are performed, how many concurrent users connect at peak time, and so on.

• Audit users, statements, or objects.

• Audit data retrieval, data manipulation, and data definition statements.

• Write the audit trail to a centralized audit table.

• Generate audit records once per session or once per access attempt.

• Capture successful attempts, unsuccessful attempts, or both.

• Enable and disable dynamically.

Executing SQL through PL/SQL program units may generate several audit records because the programunits may refer to other database objects.

Instructor NoteThe data dictionary view SYS.AUD$ contains all information regarding audit. Access the view asSYSTEM user.

AUDIT statement_opt|system_priv[, ...] [BY user[, ...]] [BY[SESSION|ACCESS]] [WHENEVER [NOT] SUCCESSFUL];

Examples:AUDIT ROLE; AUDIT ROLE WHENEVER SUCCESSFUL;

AUDIT ROLE WHENEVER NOT SUCCESSFUL;

AUDIT SELECT TABLE, UPDATE TABLE BY scott, blake;



CREATE OR REPLACE TRIGGER audit_emp_valuesAFTER DELETE OR INSERT OR UPDATE ON employeesFOR EACH ROWBEGINIF (audit_emp_package.g_reason IS NULL) THEN

RAISE_APPLICATION_ERROR (-20059, 'Specify a reasonfor the data operation through the procedure SET_REASONof the AUDIT_EMP_PACKAGE before proceeding.');

ELSEINSERT INTO audit_emp_table (user_name, timestamp, id,old_last_name, new_last_name, old_title, new_title,old_salary, new_salary, comments)

VALUES (USER, SYSDATE, :OLD.employee_id, :OLD.last_name,:NEW.last_name, :OLD.job_id, :NEW.job_id, :OLD.salary,:NEW.salary, audit_emp_package.g_reason);

END IF;END;

CREATE OR REPLACE TRIGGER cleanup_audit_empAFTER INSERT OR UPDATE OR DELETE ON employeesBEGINaudit_emp_package.g_reason := NULL;

END;

Auditing by Using a Trigger

Audit Data Values

Audit actual data values with triggers.

You can:

• Audit data manipulation statements only

• Write the audit trail to a user-defined audit table

• Generate audit records once for the statement or once for each row

• Capture successful attempts only

• Enable and disable dynamically

Using the Oracle server, you can perform database auditing. Database auditing cannot record changes tospecific column values. If the changes to the table columns need to be tracked and column values need to bestored for each change, use application auditing. Application auditing can be done either through storedprocedures or database triggers, as shown in the example in the slide.

Instructor NoteThe user makes a call to the AUDIT_EMP_PACKAGE and sets the variable G_REASON. During a DMLoperation, the trigger checks the reason for the data manipulation operation. If G_REASON is not set, an erroris raised; if it is set, the DML operation is successful, and it is recorded in the AUDIT_EMP_TABLE. If theAUDIT_EMP_TABLE is not already existing, run the script file demo\09_addtabs.sql.

The second trigger clears the variable G_REASON of a value.

The code example script file for this example contains additional statements that you can use to create thepackage used by the trigger and the statements to test the trigger.



ALTER TABLE employees ADDCONSTRAINT ck_salary CHECK (salary >= 500);

Enforcing Data IntegrityWithin the Server

Enforcing Data Integrity within the Server

You can enforce data integrity within the Oracle server and develop triggers to handle more complexdata integrity rules.

The standard data integrity rules are not null, unique, primary key, and foreign key.

Use these rules to:

• Provide constant default values

• Enforce static constraints

• Enable and disable dynamically

Example

The code sample in the slide ensures that the salary is at least $500.



CREATE OR REPLACE TRIGGER check_salaryBEFORE UPDATE OF salary ON employeesFOR EACH ROWWHEN (NEW.salary < OLD.salary)

BEGINRAISE_APPLICATION_ERROR (-20508,

'Do not decrease salary.');END;/

Protecting Data Integritywith a Trigger

Protecting Data Integrity with a Trigger

Protect data integrity with a trigger and enforce nonstandard data integrity checks.

• Provide variable default values.

• Enforce dynamic constraints.


• Incorporate declarative constraints within the definition of a table to protect data integrity.

Example

The code sample in the slide ensures that the salary is never decreased.

Instructor Note

The code example script file for this example contains additional statements to test the trigger.



ALTER TABLE employeesADD CONSTRAINT emp_deptno_fkFOREIGN KEY (department_id)

REFERENCES departments(department_id)ON DELETE CASCADE;

Enforcing Referential IntegrityWithin the Server

Enforcing Referential Integrity within the Server

Incorporate referential integrity constraints within the definition of a table to prevent data inconsistencyand enforce referential integrity within the server.

• Restrict updates and deletes.

• Cascade deletes.


Example

When a department is removed from the DEPARTMENTS parent table, cascade the deletion to thecorresponding rows in the EMPLOYEES child table.

Instructor Note

Do not run the code example for the code in the slide. The code produces an error message because theforeign key constraint mentioned in the code is already existing in the table.



CREATE OR REPLACE TRIGGER cascade_updates

AFTER UPDATE OF department_id ON departments

FOR EACH ROW

BEGIN

UPDATE employees

SET employees.department_id=:NEW.department_id

WHERE employees.department_id=:OLD.department_id;

UPDATE job_history

SET department_id=:NEW.department_id

WHERE department_id=:OLD.department_id;

END;

/

Protecting Referential Integritywith a Trigger

Protecting Referential Integrity with a Trigger

Develop triggers to implement referential integrity rules that are not supported by declarativeconstraints.

• Cascade updates.

• Set to NULL for updates and deletions.

• Set to a default value on updates and deletions.

• Enforce referential integrity in a distributed system.


Example

Enforce referential integrity with a trigger. When the value of DEPARTMENT_ID changes in theDEPARTMENTS parent table, cascade the update to the corresponding rows in the EMPLOYEES childtable.

For a complete referential integrity solution using triggers, a single trigger is not enough.

Instructor Note

The code example script file for this page contains additional statements that temporarily disableconstraints on the EMPLOYEES and JOB_HISTORY tables. The script file also tests the triggers, rollsback the transaction, and re-enables the triggers.



CREATE SNAPSHOT emp_copy ASSELECT * FROM employees@ny;

Replicating a TableWithin the Server

Creating a Snapshot

A snapshot is a local copy of a table data that originates from one or more remote master tables. Anapplication can query the data in a read-only table snapshot, but cannot insert, update, or delete rows inthe snapshot. To keep a snapshot's data current with the data of its master, the Oracle server mustperiodically refresh the snapshot.

When this statement is used in SQL, replication is performed implicitly by the Oracle server by usinginternal triggers. This has better performance over using user-defined PL/SQL triggers for replication.

Copying Tables with Server Snapshots

Copy a table with a snapshot.

• Copy tables asynchronously, at user-defined intervals.

• Base snapshots on multiple master tables.

• Read from snapshots only.

• Improve the performance of data manipulation on the master table, particularly if the network fails.

Alternatively, you can replicate tables using triggers.

Example

In San Francisco, create a snapshot of the remote EMPLOYEES table in New York.

Instructor Note

Replication of snapshots or materialized views is done within the database kernel and is faster than atrigger. There are no code example script files that you can use to demonstrate replication.



CREATE OR REPLACE TRIGGER emp_replicaBEFORE INSERT OR UPDATE ON employeesFOR EACH ROWBEGIN /*Only proceed if user initiates a data operation,

NOT through the cascading trigger.*/IF INSERTING THEN

IF :NEW.flag IS NULL THENINSERT INTO employees@sfVALUES(:new.employee_id, :new.last_name,..., 'B');:NEW.flag := 'A';

END IF;ELSE /* Updating. */IF :NEW.flag = :OLD.flag THEN

UPDATE employees@sfSET ename = :NEW.last_name, ...,

flag = :NEW.flagWHERE employee_id = :NEW.employee_id;

END IF;IF :OLD.flag = 'A' THEN :NEW.flag := 'B';ELSE :NEW.flag := 'A';END IF;

END IF;END;

Replicating a Table with a Trigger

Replicating a Table with a Trigger

Replicate a table with a trigger.

• Copy tables synchronously, in real time.

• Base replicas on a single master table.

• Read from replicas, as well as write to them.

• Impair the performance of data manipulation on the master table, particularly if the network fails.

Maintain copies of tables automatically with snapshots, particularly on remote nodes.

Example

In New York, replicate the local EMPLOYEES table to San Francisco.

Instructor Note

If the network were to fail and you used the snapshot method, the Oracle server capabilities ensure thatthe users on the main node are not affected, and that the snapshot is maintained asynchronously.

If you used the trigger method of replication, the user is not able to work because the trigger would notbe able to write to the remote database.



Computing Derived Data Within the Server

UPDATE departmentsSET total_sal=(SELECT SUM(salary)

FROM employeesWHERE employees.department_id =

departments.department_id);

Computing Derived Data within the Server

Compute derived values in a batch job.

• Compute derived column values asynchronously, at user-defined intervals.

• Store derived values only within database tables.

• Modify data in one pass to the database and calculate derived data in a second pass.

Alternatively, you can use triggers to keep running computations of derived data.

Example

Keep the salary total for each department within a special TOTAL_SALARY column of theDEPARTMENTS table.

Instructor Note

The code example script file for this topic contains additional statements that you can use to add theTOTAL_SAL column to the DEPARTMENTS table and then drop it from the table.



CREATE OR REPLACE PROCEDURE increment_salary(p_id IN departments.department_id%TYPE,p_salary IN departments.total_sal%TYPE)

ISBEGIN

UPDATE departmentsSET total_sal = NVL (total_sal, 0)+ p_salaryWHERE department_id = p_id;

END increment_salary;

CREATE OR REPLACE TRIGGER compute_salaryAFTER INSERT OR UPDATE OF salary OR DELETE ON employeesFOR EACH ROWBEGINIF DELETING THEN

increment_salary(:OLD.department_id,(-1*:OLD.salary));ELSIF UPDATING THENincrement_salary(:NEW.department_id,(:NEW.salary-:OLD.salary));ELSE increment_salary(:NEW.department_id,:NEW.salary);--INSERTEND IF;END;

Computing Derived Values with a Trigger

Computing Derived Data Values with a Trigger

Compute derived values with a trigger.

• Compute derived columns synchronously, in real time.

• Store derived values within database tables or within package global variables.

• Modify data and calculate derived data in a single pass to the database.

Example

Keep a running total of the salary for each department within the special TOTAL_SALARY column ofthe DEPARTMENTS table.

Instructor Note

The code example script file for this topic contains additional statements that you can use to add theTOTAL_SAL column to the DEPARTMENTS table and then drop it from the table. The script alsocontains statements to test the trigger and roll back the transaction at the end.



CREATE OR REPLACE TRIGGER notify_reorder_repBEFORE UPDATE OF quantity_on_hand, reorder_pointON inventories FOR EACH ROWDECLAREv_descrip product_descriptions.product_description%TYPE;v_msg_text VARCHAR2(2000);stat_send number(1);BEGINIF :NEW.quantity_on_hand <= :NEW.reorder_point THENSELECT product_description INTO v_descripFROM product_descriptionsWHERE product_id = :NEW.product_id;v_msg_text := 'ALERT: INVENTORY LOW ORDER:'||CHR(10)||...'Yours,' ||CHR(10) ||user || '.'|| CHR(10)|| CHR(10);ELSIF:OLD.quantity_on_hand < :NEW.quantity_on_hand THEN NULL;ELSEv_msg_text := 'Product #'||... CHR(10);END IF;DBMS_PIPE.PACK_MESSAGE(v_msg_text);stat_send := DBMS_PIPE.SEND_MESSAGE('INV_PIPE');

END;

Logging Events with a Trigger

Logging Events with a Trigger

Within the server, you can log events by querying data and performing operations manually. This sendsa message using a pipe when the inventory for a particular product has fallen below the acceptablelimit. This trigger uses the Oracle-supplied package DBMS_PIPE to send the message.

Logging Events within the Server

• Query data explicitly to determine whether an operation is necessary.

• In a second step, perform the operation, such as sending a message.

Using Triggers to Log Events

• Perform operations implicitly, such as firing off an automatic electronic memo.

• Modify data and perform its dependent operation in a single step.

• Log events automatically as data is changing.

Instructor Note

There are no code example script files for logging events with triggers.


Logging Events with a Trigger (continued)

Logging Events Transparently

In the trigger code:

• CHR(10) is a carriage return

• Reorder_point is not null

• Another transaction can receive and read the message in the pipe

ExampleCREATE OR REPLACE TRIGGER notify_reorder_rep

BEFORE UPDATE OF amount_in_stock, reorder_point

ON inventory FOR EACH ROWDECLARE

v_descrip product.descrip%TYPE;

v_msg_text VARCHAR2(2000);

stat_send number(1);

BEGIN

IF :NEW.amount_in_stock <= :NEW.reorder_point THEN

SELECT descrip INTO v_descrip

FROM PRODUCT WHERE prodid = :NEW.product_id;

v_msg_text := 'ALERT: INVENTORY LOW ORDER:'||CHR(10)||

'It has come to my personal attention that, due to recent'

||CHR(10)||'transactions, our inventory for product # '||

TO_CHAR(:NEW.product_id)||'-- '||v_descrip ||

' -- has fallen below acceptable levels.' || CHR(10) ||

'Yours,' ||CHR(10) ||user || '.'|| CHR(10)|| CHR(10);

ELSIF

:OLD.amount_in_stock<:NEW.amount_in_stock THEN NULL;ELSE

v_msg_text := 'Product #'|| TO_CHAR(:NEW.product_id)

||' ordered. '|| CHR(10)|| CHR(10); END IF;

DBMS_PIPE.PACK_MESSAGE(v_msg_text);

stat_send := DBMS_PIPE.SEND_MESSAGE('INV_PIPE');

END;



Benefits of Database Triggers

• Improved data security:

– Provide enhanced and complex securitychecks

– Provide enhanced and complex auditing

• Improved data integrity:

– Enforce dynamic data integrity constraints

– Enforce complex referential integrityconstraints

– Ensure that related operations are performedtogether implicitly

Benefits of Database Triggers

You can use database triggers:

• As alternatives to features provided by the Oracle server

• If your requirements are more complex or more simple than those provided by the Oracle server

• If your requirements are not provided by the Oracle server at all



Managing Triggers

The following system privileges are required tomanage triggers:• The CREATE/ALTER/DROP (ANY) TRIGGER

privilege enables you to create a trigger in anyschema

• The ADMINISTER DATABASE TRIGGER privilegeenables you to create a trigger on DATABASE

• The EXECUTE privilege (if your trigger refers to anyobjects that are not in your schema)

Note: Statements in the trigger body operate underthe privilege of the trigger owner, not the trigger user.

Managing TriggersIn order to create a trigger in your schema, you need the CREATE TRIGGER system privilege, andyou must either own the table specified in the triggering statement, have the ALTER privilege forthe table in the triggering statement, or have the ALTER ANY TABLE system privilege. You canalter or drop your triggers without any further privileges being required.

If the ANY keyword is used, you can create, alter, or drop your own triggers and those in anotherschema and can be associated with any user’s table.

You do not need any privileges to invoke a trigger in your schema. A trigger is invoked by DMLstatements that you issue. But if your trigger refers to any objects that are not in your schema, theuser creating the trigger must have the EXECUTE privilege on the referenced procedures, functions,or packages, and not through roles. As with stored procedures, the statement in the trigger bodyoperates under the privilege domain of the trigger’s owner, not that of the user issuing thetriggering statement.

To create a trigger on DATABASE, you must have the ADMINISTER DATABASE TRIGGERprivilege. If this privilege is later revoked, you can drop the trigger, but you cannot alter it.



Viewing Trigger Information

You can view the following trigger information:• USER_OBJECTS data dictionary view: object

information

• USER_TRIGGERS data dictionary view: the text ofthe trigger

• USER_ERRORS data dictionary view: PL/SQL syntaxerrors (compilation errors) of the trigger

Viewing Trigger Information

The slide shows the data dictionary views that you can access to get information regarding thetriggers.

The USER_OBJECTS view contains the name and status of the trigger and the date and time whenthe trigger was created.

The USER_ERRORS view contains the details of the compilation errors that occurred while atrigger was compiling. The contents of these views are similar to those for subprograms.

The USER_TRIGGERS view contains details such as name, type, triggering event, the table onwhich the trigger is created, and the body of the trigger.

The SELECT Username FROM USER_USERS; statement gives the name of the owner of thetrigger, not the name of the user who is updating the table.



Column

TRIGGER_NAME

TRIGGER_TYPE

TRIGGERING_EVENT

TABLE_NAME

REFERENCING_NAMES

WHEN_CLAUSE

STATUS

TRIGGER_BODY

Column Description

Name of the trigger

The type is BEFORE, AFTER, INSTEAD OF

The DML operation firing the trigger

Name of the database table

Name used for :OLD and :NEW

The when_clause used

The status of the trigger

The action to take

Using USER_TRIGGERS

Abridged column list*

*

Using USER_TRIGGERS

If the source file is unavailable, you can use iSQL*Plus to regenerate it from USER_TRIGGERS.You can also examine the ALL_TRIGGERS and DBA_TRIGGERS views, each of which contains theadditional column OWNER, for the owner of the object.

Instructor NoteThe TRIGGER_BODY column of the USER_TRIGGERS data dictionary view is of the LONG datatype. To view the data in this column, you may need to use the SET LONG option in iSQL*Plus toincrease the number of characters displayed. By default, the value of the LONG data type is set to 80.



Listing the Code of Triggers

SELECT trigger_name, trigger_type, triggering_event,table_name, referencing_names,status, trigger_body

FROM user_triggersWHERE trigger_name = 'RESTRICT_SALARY';

ExampleUse the USER_TRIGGERS data dictionary view to display information about the RESTRICT_SALtrigger.



Summary


• Use advanced database triggers

• List mutating and constraining rules for triggers

• Describe the real-world application of triggers

• Manage triggers

• View trigger information




This practice covers creating advanced triggers toadd to the capabilities of the Oracle database.


In this practice you decide how to implement a number of business rules. You will create triggers forthose rules that should be implemented as triggers. The triggers will execute procedures that youhave placed in a package.

Instructor NoteIf you have created the CHECK_SALARY trigger, you need to drop it if you want to do some of thepractices in this lesson and the following lessons. It conflicts with a trigger that is created in thepractices.


Practice 17A number of business rules that apply to the EMPLOYEES and DEPARTMENTS tables are listed below.Decide how to implement each of these business rules, by means of declarative constraints or by usingtriggers.Which constraints or triggers are needed and are there any problems to be expected?Implement the business rules by defining the triggers or constraints that you decided to create.A partial package is provided in file lab17_1.sql to which you should add any necessary proceduresor functions that are to be called from triggers that you may create for the following rules.(The triggers should execute procedures or functions that you have defined in the package.)

Business Rules

Rule 1. Sales managers and sales representatives should always receive commission. Employeeswho are not sales managers or sales representatives should never receive a commission.Ensure that this restriction does not validate the existing records of the EMPLOYEEStable. It should be effective only for the subsequent inserts and updates on the table.

Rule 2. The EMPLOYEES table should contain exactly one president.

Test your answer by inserting an employee record with the following details: employeeID 400, last name Harris, first name Alice, e-mail ID AHARRIS, job ID AD_PRES,hire date SYSDATE , salary 20000, and department ID 20.

Note: You do not need to implement a rule for case sensitivity; instead you need to testfor the number of people with the job title of President.

Rule 3. An employee should never be a manager of more than 15 employees.

Test your answer by inserting the following records into the EMPLOYEES table (performa query to count the number of employees currently working for manager 100 beforeinserting these rows):

i. Employee ID 401, last name Johnson, first name Brian, e-mail IDBJOHNSON, job ID SA_MAN, hire date SYSDATE, salary 11000, managerID 100, and department ID 80. (This insertion should be successful, becausethere are only 14 employees working for manager 100 so far.)

ii. Employee ID 402, last name Kellogg, first name Tony, e-mail IDTKELLOG, job ID ST_MAN, hire date SYSDATE , salary 7500, managerID 100, and department ID 50. (This insertion should be unsuccessful, because thereare already 15 employees working for manager 100.)

Rule 4. Salaries can only be increased, never decreased.

The present salary of employee 105 is 5000. Test your answer by decreasing the salary ofemployee 105 to 4500.


Practice 17 (continued)

Rule 5. If a department moves to another location, each employee of that department automaticallyreceives a salary raise of 2 percent.

View the current salaries of employees in department 90.

Test your answer by moving department 90 to location 1600. Query the new salaries ofemployees of department 90.


Managing Dependencies


40 minutes Practice

85 minutes Total



Objectives


• Track procedural dependencies

• Predict the effect of changing a database objectupon stored procedures and functions

• Manage procedural dependencies

Lesson Aim

This lesson introduces you to object dependencies and implicit and explicit recompilation of invalidobjects.



Understanding Dependencies

Table

View

Database Trigger

Procedure

Function

Package Body

Package Specification

User-Defined Objectand Collection Types

Function

Package Specification

Procedure

Sequence

Synonym

Table

View

User-Defined Objectand Collection Types

Referenced ObjectsDependent Objects

Dependent and Referenced Objects

Some objects reference other objects as part of their definition. For example, a stored procedurecould contain a SELECT statement that selects columns from a table. For this reason, the storedprocedure is called a dependent object, whereas the table is called a referenced object.

Dependency Issues

If you alter the definition of a referenced object, dependent objects may or may not continue towork properly. For example, if the table definition is changed, the procedure may or may notcontinue to work without error.

The Oracle server automatically records dependencies among objects. To manage dependencies, allschema objects have a status (valid or invalid) that is recorded in the data dictionary, and you canview the status in the USER_OBJECTS data dictionary view.

Status Significance

VALID The schema object has been compiled and can be immediately used whenreferenced.

INVALID The schema object must be compiled before it can be used.



Dependencies

xxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvv

ProcedureView or

procedure

Directdependency

Dependent

Referenced

Indirectdependency

Referenced

Directdependency

Dependent

Table

Referenced

Dependent and Referenced Objects (continued)

A procedure or a function can directly or indirectly (through an intermediate view,procedure, function, or packaged procedure or function) reference the following objects:

• Tables

• Views

• Sequences

• Procedures

• Functions

• Packaged procedures or functions



Local Dependencies

xxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvvvvvvvvvvv

Procedure View

Local references

Procedure Tablevvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvv

Direct localdependency

Managing Local Dependencies

In the case of local dependencies, the objects are on the same node in the same database. The Oracleserver automatically manages all local dependencies, using the database’s internal “depends-on”table. When a referenced object is modified, the dependent objects are invalidated. The next time aninvalidated object is called, the Oracle server automatically recompiles it.



Local Dependencies

xxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvvvvvvvvvvv

Procedure View

Local references

Procedure Tablevvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvv


Definitionchange

INVALIDINVALIDINVALID

The Oracle server implicitly recompiles any INVALIDobject when the object is next called.

Managing Local Dependencies (continued)

Assume that the structure of the table on which a view is based is modified. When you describe theview by using iSQL*Plus DESCRIBE command, you get an error message that states that the objectis invalid to describe. This is because the command is not a SQL command and, at this stage, theview is invalid because the structure of its base table is changed. If you query the view now, theview is recompiled automatically and you can see the result if it is successfully recompiled.



xxxxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxvvvvvvvvvvvvvvvvvvvvv


ADD_EMPprocedure

QUERY_EMPprocedure

EMPLOYEES table

EMP_VW view

A Scenario of Local Dependencies

…

…

ExampleThe QUERY_EMP procedure directly references the EMPLOYEES table. The ADD_EMP procedureupdates the EMPLOYEES table indirectly, by way of the EMP_VW view.

In each of the following cases, will the ADD_EMP procedure be invalidated, and will itsuccessfully recompile?

1. The internal logic of the QUERY_EMP procedure is modified.

2. A new column is added to the EMPLOYEES table.

3. The EMP_VW view is dropped.

Instructor NoteRun the script file 18_localdep.sql to create view and procedures shown in the slide.

Ask the class as a whole to answer the questions and discuss the various answers that are given bythe participants. Recompilation is discussed in more detail later in this lesson.

Demonstrate the responses if you have time and wish to do so.

Answer to case 1: Modifying QUERY_EMP has no effect on ADD_EMP.

Answer to case 2: ADD_EMP will be invalidated, and it will successfully recompile when invoked.This is true if a column list is provided with all INSERT statements in the procedure, and if thenew column that is being added to the table does not have a NOT NULL constraint.

Answer to case 3: ADD_EMP will be invalidated, and it will never compile successfully as long asthe view is not back in place.



Displaying Direct Dependencies by UsingUSER_DEPENDENCIES

SELECT name, type, referenced_name, referenced_typeFROM user_dependenciesWHERE referenced_name IN ('EMPLOYEES','EMP_VW' );

……

Display Direct Dependencies by Using USER_DEPENDENCIES

Determine which database objects to recompile manually by displaying direct dependencies from theUSER_DEPENDENCIES data dictionary view.

Examine the ALL_DEPENDENCIES and DBA_DEPENDENCIES views, each of which contains theadditional column OWNER, that reference the owner of the object.

Instructor Note

You may see more objects in this lesson than those created by students. This happens because youcreated more objects than the student, by running the code examples and demos in the previous lessons.You may highlight only those objects that the students created in the practices.

Column Column Description

NAME The name of the dependent object

TYPE The type of the dependent object (PROCEDURE, FUNCTION,PACKAGE, PACKAGE BODY, TRIGGER, or VIEW)

REFERENCED_OWNER The schema of the referenced object

REFERENCED_NAME The name of the referenced object

REFERENCED_TYPE The type of the referenced object

REFERENCED_LINK_NAME The database link used to access the referenced object



Displaying Direct and IndirectDependencies

1. Run the script utldtree.sql that creates theobjects that enable you to display the direct andindirect dependencies.

2. Execute the DEPTREE_FILL procedure.

EXECUTE deptree_fill('TABLE','SCOTT','EMPLOYEES')

Displaying Direct and Indirect Dependencies by Using Views Provided by OracleDisplay direct and indirect dependencies from additional user views called DEPTREE andIDEPTREE; these view are provided by Oracle.

Example

1. Make sure the utldtree.sql script has been executed. This script is located in the$ORACLE_HOME/rdbms/admin folder. (This script is supplied in the lab folder ofyour class files.)

2. Populate the DEPTREE_TEMPTAB table with information for a particular referenced objectby invoking the DEPTREE_FILL procedure. There are three parameters for this procedure:

Instructor NoteThe code example script file for the example above uses TEACH instead of SCOTT. Replace itwith your username if you are using a different schema.

object_type Is the type of the referenced object

object_owner Is the schema of the referenced object

object_name Is the name of the referenced object



Displaying Dependencies

DEPTREE View

SELECT nested_level, type, nameFROM deptreeORDER BY seq#;

…

…

ExampleDisplay a tabular representation of all dependent objects by querying the DEPTREE view.

Display an indented representation of the same information by querying the IDEPTREE view, whichconsists of a single column named DEPENDENCIES.

For example,SELECT *FROM ideptree;

provides a single column of indented output of the dependencies in a hierarchical structure.

Instructor NoteThe IDEPTREE view shows indented output only with SQL*Plus. With iSQL*Plus, the output isshown as a single column with the dependent objects following in a hierarchy.



EMPLOYEES table

REDUCE_SALprocedure

RAISE_SALprocedure

Another Scenario of Local Dependencies



…

Predicting the Effects of Changes on Dependent Objects

Example 1

Predict the effect that a change in the definition of a procedure has on the recompilation of adependent procedure.

Suppose that the RAISE_SAL procedure updates the EMPLOYEES table directly, and that theREDUCE_SAL procedure updates the EMPLOYEES table indirectly by way of RAISE_SAL.

In each of the following cases, will the REDUCE_SAL procedure successfully recompile?

1. The internal logic of the RAISE_SAL procedure is modified.

2. One of the formal parameters to the RAISE_SAL procedure is eliminated.

Instructor Note

Pose the questions above to the class as a whole and discuss the various answers that are given bythe participants. There is more information on recompilation later in this lesson.

Do a demonstration if you have time and wish to do so.

Answer to case 1: REDUCE_SAL will successfully recompile if RAISE_SAL has successfullycompiled.

Answer to case 2: REDUCE_SAL will not successfully recompile.



QUERY_EMPprocedure EMPLOYEES public synonym

X

A Scenario of Local NamingDependencies


EMPLOYEEStable

…

…

Predicting Effects of Changes on Dependent Objects (continued)

Example 2Be aware of the subtle case in which the creation of a table, view, or synonym may unexpectedlyinvalidate a dependent object because it interferes with the Oracle server hierarchy for resolvingname references.Predict the effect that the name of a new object has upon a dependent procedure.Suppose that your QUERY_EMP procedure originally referenced a public synonym calledEMPLOYEES. However, you have just created a new table called EMPLOYEES within your ownschema. Will this change invalidate the procedure? Which of the two EMPLOYEES objects willQUERY_EMP reference when the procedure recompiles?Now suppose that you drop your private EMPLOYEES table. Will this invalidate the procedure?What will happen when the procedure recompiles?You can track security dependencies within the USER_TAB_PRIVS data dictionary view.

Instructor NoteIf your EMPLOYEES table is a test copy of the public synonym EMPLOYEES table, then, yes,recompilation of QUERY_EMP will be successful and will reference your test EMPLOYEES table. Ifyour EMPLOYEES table has different column names, then QUERY_EMP gives compile errors andremains invalid.Also, if you drop your private EMPLOYEES table, your dependent objects become invalid. If youhave access to a public object called EMPLOYEES through a public synonym, that has the samestructure, the recompiled objects now refer to the public EMPLOYEES object.




Procedure ViewProcedure Table

vvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvvxxxxxxxxxxxxxxvvvvvvvvvvvvvv


Direct remotedependency

Understanding Remote Dependencies

Local and remote references

Network


In the case of remote dependencies, the objects are on separate nodes. The Oracle server does notmanage dependencies among remote schema objects other than local-procedure-to-remote-procedure dependencies (including functions, packages, and triggers). The local stored procedureand all of its dependent objects will be invalidated but will not automatically recompile when calledfor the first time.







Direct remotedependency


Local and remote references

Definitionchange

INVALIDINVALIDVALID

Network

Understanding Remote Dependencies (continued)

Recompilation of Dependent Objects: Local and Remote

• Verify successful explicit recompilation of the dependent remote procedures and implicitrecompilation of the dependent local procedures by checking the status of these procedureswithin the USER_OBJECTS view.

• If an automatic implicit recompilation of the dependent local procedures fails, the statusremains invalid and the Oracle server issues a run-time error. Therefore, to avoid disruptingproduction, it is strongly recommended that you recompile local dependent objects manually,rather than relying on an automatic mechanism.



Concepts of Remote Dependencies

Remote dependencies are governed by the modechosen by the user:

• TIMESTAMP checking

• SIGNATURE checking

TIMESTAMP Checking

Each PL/SQL program unit carries a time stamp that is set when it is created or recompiled.Whenever you alter a PL/SQL program unit or a relevant schema object, all of its dependentprogram units are marked as invalid and must be recompiled before they can execute. The actualtime stamp comparison occurs when a statement in the body of a local procedure calls a remoteprocedure.

SIGNATURE Checking

For each PL/SQL program unit, both the time stamp and the signature are recorded. The signatureof a PL/SQL construct contains information about the following:

• The name of the construct (procedure, function, or package)

• The base types of the parameters of the construct

• The modes of the parameters (IN, OUT, or IN OUT)

• The number of the parameters

The recorded time stamp in the calling program unit is compared with the current time stamp in thecalled remote program unit. If the time stamps match, the call proceeds normally. If they do notmatch, the Remote Procedure Calls (RPC) layer performs a simple test to compare the signature todetermine whether the call is safe or not. If the signature has not been changed in an incompatiblemanner, execution continues; otherwise, an error status is returned.



REMOTE_DEPENDENCIES_MODE Parameter

Setting REMOTE_DEPENDENCIES_MODE:

• As an init.ora parameter

REMOTE_DEPENDENCIES_MODE = value

• At the system level

ALTER SYSTEM SETREMOTE_DEPENDENCIES_MODE = value

• At the session level

ALTER SESSION SETREMOTE_DEPENDENCIES_MODE = value

Setting the REMOTE_DEPENDENCIES_MODE

value TIMESTAMPSIGNATURE

Specify the value of the REMOTE_DEPENDENCIES_MODE parameter, using one of the threemethods described in the slide.

Note: The calling site determines the dependency model.



Remote Dependencies andTime Stamp Mode




Network

Network

Using Time Stamp Mode for Automatic Recompilation of Local and Remote Objects

If time stamps are used to handle dependencies among PL/SQL program units, then whenever youalter a program unit or a relevant schema object, all of its dependent units are marked as invalid andmust be recompiled before they can be run.



Remote Dependencies andTime Stamp Mode




Network

Network Definitionchange

INVALIDINVALIDVALID

Using Time Stamp Mode for Automatic Recompilation of Local and Remote Objects

In the example in the slide, the definition of the table changes. Hence, all of its dependent units aremarked as invalid and must be recompiled before they can be run.

• When remote objects change, it is strongly recommended that you recompile local dependentobjects manually in order to avoid disrupting production.

• The remote dependency mechanism is different from the automatic local dependencymechanism already discussed. The first time a recompiled remote subprogram is invoked by alocal subprogram, you get an execution error and the local subprogram is invalidated; thesecond time it is invoked, implicit automatic recompilation takes place.



Remote Procedure B Compilesat 8:00 a.m.

Valid

Remote procedure B

Compiles

Local Procedures Referencing Remote Procedures

A local procedure that references a remote procedure is invalidated by the Oracle server if theremote procedure is recompiled after the local procedure is compiled.

Automatic Remote Dependency Mechanism

When a procedure compiles, the Oracle server records the time stamp of that compilation withinthe P code of the procedure.

In the slide, when the remote procedure B was successfully compiled at 8 a.m., this time wasrecorded as its time stamp



Local Procedure A Compilesat 9:00 a.m.

Local procedure A

Valid

Remote procedure B

Time stampof B

Valid

Time stampof A

RecordTime stampof B

Automatic Remote Dependency Mechanism

When a local procedure referencing a remote procedure compiles, the Oracle server also records the timestamp of the remote procedure into the P code of the local procedure.

In the slide, local procedure A which is dependent on remote procedure B is compiled at 9:00 a.m. Thetime stamps of both procedure A and remote procedure B are recorded in the P code of procedure A.



Execute Procedure A

Local procedure A

Valid

Remote procedure B

Time stampof B

Valid

Time stampof A

Time stampof B

Time stampcomparison

Execute B

Automatic Remote Dependency

When the local procedure is invoked, at run time the Oracle server compares the two time stampsof the referenced remote procedure.

If the time stamps are equal (indicating that the remote procedure has not recompiled), the Oracleserver executes the local procedure.

In the example in the slide, the time stamp recorded with P code of remote procedure B is the sameas that recorded with local procedure A. Hence, local procedure A is valid.



Remote Procedure B Recompiledat 11:00 a.m.

Valid

Remote procedure B

Compiles

Local Procedures Referencing Remote Procedures

Assume that the remote procedure B is successfully recompiled at 11a.m. The new time stamp isrecorded along with its P code.



Execute Procedure A

Local procedure A

Valid

Remote procedure B

Time stampof B

Valid

Time stampof A

Time stampof B

Time stampcomparison

ERROR

Invalid

Automatic Remote Dependency

If the time stamps are not equal (indicating that the remote procedure has recompiled), the Oracleserver invalidates the local procedure and returns a runtime error.

If the local procedure, which is now tagged as invalid, is invoked a second time, the Oracle serverrecompiles it before executing, in accordance with the automatic local dependency mechanism.

Note: If a local procedure returns a run-time error the first time that it is invoked, indicating that theremote procedure's time stamp has changed, you should develop a strategy to reinvoke the localprocedure.

In the example in the slide, remote procedure is recompiled at 11 a.m. and this time is recorded asits time stamp in the P code. The P code of local procedure A still has 8 a.m. as time stamp for theremote procedure B.

Because the time stamp recorded with P code of local procedure A is different from that recordedwith remote procedure B, the local procedure is marked invalid. When the local procedure isinvoked for the second time, it can be successfully compiled and marked valid.

Disadvantage of time stamp mode

A disadvantage of the time stamp mode is that it is unnecessarily restrictive. Recompilation ofdependent objects across the network are often performed when not strictly necessary, leading toperformance degradation.

Instructor NoteDiscuss the scenario in the slide, assuming that the calling side uses SIGNATURE mode.



Signature Mode

• The signature of a procedure is:

– The name of the procedure

– The datatypes of the parameters

– The modes of the parameters

• The signature of the remote procedure is saved inthe local procedure.

• When executing a dependent procedure, thesignature of the referenced remote procedure iscompared.

Signatures

To alleviate some of the problems with the time stamp-only dependency model, you can use thesignature model. This allows the remote procedure to be recompiled without affecting the localprocedures. This is important if the database is distributed.

The signature of a subprogram contains the following information:

• The name of the subprogram

• The datatypes of the parameters

• The modes of the parameters

• The number of parameters

• The datatype of the return value for a function

If a remote program is changed and recompiled but the signature does not change, then the localprocedure can execute the remote procedure. With the time stamp method, an error would have beenraised because the time stamps would not have matched.


Recompiling PL/SQL ObjectsIf the recompilation is successful, the object becomes valid. If not, the Oracle server returns an errorand the object remains invalid.When you recompile a PL/SQL object, the Oracle server first recompiles any invalid objects on whichit depends.ProcedureAny local objects that depend on a procedure (such as procedures that call the recompiled procedure orpackage bodies that define the procedures that call the recompiled procedure) are also invalidated.PackagesThe COMPILE PACKAGE option recompiles both the package specification and the body, regardless ofwhether it is invalid. The COMPILE BODY option recompiles only the package body.Recompiling a package specification invalidates any local objects that depend on the specification, suchas procedures that call procedures or functions in the package. Note that the body of a package alsodepends on its specification.TriggersExplicit recompilation eliminates the need for implicit run-time recompilation and prevents associatedrun-time compilation errors and performance overhead.The DEBUG option instructs the PL/SQL compiler to generate and store the code for use by thePL/SQL debugger.

Instructor NoteExplicit compiling of subprograms can also be done programmatically, using the DBMS_DDL orDBMS_UTILITY packages.


Recompiling a PL/SQLProgram Unit

Recompilation:

• Is handled automatically through implicit run-timerecompilation

• Is handled through explicit recompilation with theALTER statement

ALTER PROCEDURE [SCHEMA.]procedure_name COMPILE;

ALTER FUNCTION [SCHEMA.]function_name COMPILE;

ALTER PACKAGE [SCHEMA.]package_name COMPILE [PACKAGE];ALTER PACKAGE [SCHEMA.]package_name COMPILE BODY;

ALTER TRIGGER trigger_name [COMPILE[DEBUG]];



Unsuccessful Recompilation

Recompiling dependent procedures and functions isunsuccessful when:

• The referenced object is dropped or renamed

• The data type of the referenced column is changed

• The referenced column is dropped

• A referenced view is replaced by a view withdifferent columns

• The parameter list of a referenced procedure ismodified

Unsuccessful Recompilation

Sometimes a recompilation of dependent procedures is unsuccessful, for example, when a referencedtable is dropped or renamed.

The success of any recompilation is based on the exact dependency. If a referenced view is recreated,any object that is dependent on the view needs to be recompiled. The success of the recompilationdepends on the columns that the view now contains, as well as the columns that the dependent objectsrequire for their execution. If the required columns are not part of the new view, the object remainsinvalid.

Instructor NoteThe DROP COLUMN was introduced in Oracle8i.



Successful Recompilation

Recompiling dependent procedures and functions issuccessful if:

• The referenced table has new columns

• The data type of referenced columns has notchanged

• A private table is dropped, but a public table,having the same name and structure, exists

• The PL/SQL body of a referenced procedure hasbeen modified and recompiled successfully

Successful Recompilation

The recompilation of dependent objects is successful if:

• New columns are added to a referenced table

• All INSERT statements include a column list

• No new column is defined as NOT NULL

When a private table is referenced by a dependent procedure, and the private table is dropped, thestatus of the dependent procedure becomes invalid. When the procedure is recompiled, eitherexplicitly or implicitly, and a public table exists, the procedure can recompile successfully but is nowdependent on the public table. The recompilation is successful only if the public table contains thecolumns that the procedure requires; otherwise, the status of the procedure remains invalid.



Recompilation of Procedures

Minimize dependency failures by:

• Declaring records by using the %ROWTYPE attribute

• Declaring variables with the %TYPE attribute

• Querying with the SELECT * notation

• Including a column list with INSERT statements

Recompilation of Procedures

You can minimize recompilation failure by following the guidelines that are shown in the slide.


Managing Dependencies

You can greatly simplify dependency management with packages when referencing a packageprocedure or function from a stand-alone procedure or function.

• If the package body changes and the package specification does not change, the stand-aloneprocedure referencing a package construct remains valid.

• If the package specification changes, the outside procedure referencing a package construct isinvalidated, as is the package body.


Packages and Dependencies


Package specification

Package body


Stand-alone

procedure

Valid

Valid

Definition changed


Managing Dependencies (continued)

If a stand-alone procedure referenced within the package changes, the entire package body isinvalidated, but the package specification remains valid. Therefore, it is recommended that you bringthe procedure into the package.


Packages and Dependencies


Package specification

Package body


Stand-aloneprocedure

Invalid

Valid

Definitionchanged



Summary


• Keep track of dependent procedures

• Recompile procedures manually as soon aspossible after the definition of a database objectchanges

Lesson Summary

Avoid disrupting production by keeping track of dependent procedures and recompiling themmanually as soon as possible after the definition of a database object changes.

Situation Automatic Recompilation

Procedure depends on a local object Yes, at first re-execution

Procedure depends on a remote procedure Yes, but at second re-execution; use manualrecompilation for first re-execution, or reinvoke itsecond time

Procedure depends on a remote object otherthan a procedure

No





• Using DEPTREE_FILL and IDEPTREE to viewdependencies

• Recompiling procedures, functions, and packages

Practice 18 OverviewIn this practice you use the DEPTREE_FILL procedure and the IDEPTREE view to investigatedependencies in your schema.

In addition, you recompile invalid procedures, functions, packages, and views.


Practice 18

1. Answer the following questions.

a. Can a table or a synonym be invalid?

b. Assuming the following scenario, is the stand-alone procedure MY_PROC invalidated?

The stand-alone procedure MY_PROC depends on the packaged procedureMY_PROC_PACK.

The MY_PROC_PACK procedure’s definition is changed by recompiling the packagebody.

The MY_PROC_PACK procedure’s declaration is not altered in the package specification.

2. Execute the utldtree.sql script. This script is available in your lab folder. Print a treestructure showing all dependencies involving your NEW_EMP procedure and yourVALID_DEPTID function.

Query the IDEPTREE view to see your results. (NEW_EMP and VALID_DEPTID were created inlesson 10, “Creating Functions.” You can run the solution scripts for the practice if you need tocreate the procedure and function.)

If you have time:

3. Dynamically validate invalid objects.

a. Make a copy of your EMPLOYEES table, called EMP_COP.

b. Alter your EMPLOYEES table and add the column TOTSAL with data type NUMBER(9,2).

c. Create a script file to print the name, type, and status of all objects that are invalid.

d. Create a procedure called COMPILE_OBJ that recompiles all invalid procedures, functions,and packages in your schema.

Make use of the ALTER_COMPILE procedure in the DBMS_DDL package.

Execute the COMPILE_OBJ procedure.

e. Run the script file that you created in question 3c again and check the status column value.

Do you still have INVALID objects? If you do, why are they INVALID?

Practice Solutions

A

Oracle9i: Program with PL/SQL A-2

Practice 1 Solutions

1. Evaluate each of the following declarations. Determine which of them are not legal and explainwhy.

a. DECLAREv_id NUMBER(4);

Legal

b. DECLAREv_x, v_y, v_z VARCHAR2(10);

Illegal because only one identifier per declaration is allowed.

c. DECLAREv_birthdate DATE NOT NULL;

Illegal because the NOT NULL variable must be initialized.

d. DECLAREv_in_stock BOOLEAN := 1;

Illegal because 1 is not a Boolean expression.PL/SQL returns the following error:PLS-00382: expression is of wrong type


Practice 1 Solutions (continued)

2. In each of the following assignments, indicate whether the statement is valid and what the valid datatype of the result will be.

a. v_days_to_go := v_due_date - SYSDATE;

Valid; Number

b.v_sender := USER || ': ' || TO_CHAR(v_dept_no);

Valid; Character string

c. v_sum := $100,000 + $250,000;

Illegal; PL/SQL cannot convert special symbols from VARCHAR2 to NUMBER.

d. v_flag := TRUE;

Valid; Boolean

e. v_n1 := v_n2 > (2 * v_n3);

Valid; Boolean

f. v_value := NULL;

Valid; Any scalar data type

3. Create an anonymous block to output the phrase “My PL/SQL Block Works” to the screen.

VARIABLE g_message VARCHAR2(30)BEGIN

:g_message := 'My PL/SQL Block Works';END;/PRINT g_message

Alternate Solution:

SET SERVEROUTPUT ON

BEGIN

DBMS_OUTPUT.PUT_LINE('My PL/SQL Block Works');

END;

/



If you have time, complete the following exercise:

4. Create a block that declares two variables. Assign the value of these PL/SQL variables toiSQL*Plus host variables and print the results of the PL/SQL variables to the screen. Executeyour PL/SQL block. Save your PL/SQL block in a file named p1q4.sql, by clicking the SaveScript button. Remember to save the script with a .sql extension.

V_CHAR Character (variable length)

V_NUM Number

Assign values to these variables as follows:

Variable Value

---------- ----------------------------------------

V_CHAR The literal '42 is the answer'

V_NUM The first two characters from V_CHAR

VARIABLE g_char VARCHAR2(30)VARIABLE g_num NUMBERDECLAREv_char VARCHAR2(30);v_num NUMBER(11,2);

BEGINv_char := '42 is the answer';v_num := TO_NUMBER(SUBSTR(v_char,1,2));:g_char := v_char;:g_num := v_num;

END;/PRINT g_charPRINT g_num



DECLARE

v_weight NUMBER(3) := 600;

v_message VARCHAR2(255) := 'Product 10012';

BEGIN

/*SUBBLOCK*/

DECLARE

v_weight NUMBER(3) := 1;

v_message VARCHAR2(255) := 'Product 11001';

v_new_locn VARCHAR2(50) := 'Europe';

BEGIN

v_weight := v_weight + 1;

v_new_locn := 'Western ' || v_new_locn;

END;

v_weight := v_weight + 1;

v_message := v_message || ' is in stock';

v_new_locn := 'Western ' || v_new_locn;

END;

/

1. Evaluate the PL/SQL block above and determine the data type and value of each of the followingvariables according to the rules of scoping.

a. The value of V_WEIGHT at position 1 is:

2

The data type is NUMBER.

b. The value of V_NEW_LOCN at position 1 is:

Western Europe

The data type is VARCHAR2.

c. The value of V_WEIGHT at position 2 is:

601


d. The value of V_MESSAGE at position 2 is:

Product 10012 is in stock


e. The value of V_NEW_LOCN at position 2 is:

Illegal because v_new_locn is not visible outside the subblock.

2

1



Scope Example

DECLARE

v_customer VARCHAR2(50) := 'Womansport';

v_credit_rating VARCHAR2(50) := 'EXCELLENT';

BEGIN

DECLARE

v_customer NUMBER(7) := 201;

v_name VARCHAR2(25) := 'Unisports';

BEGIN

v_customer v_name v_credit_rating

END;

v_customer v_name v_credit_rating

END;

/



2. Suppose you embed a subblock within a block, as shown on the previous page. You declare twovariables, V_CUSTOMER and V_CREDIT_RATING, in the main block. You also declare two variables,V_CUSTOMER and V_NAME, in the subblock. Determine the values and data types for each of thefollowing cases.

a. The value of V_CUSTOMER in the subblock is:

201


b. The value of V_NAME in the subblock is:

Unisports and


c. The value of V_CREDIT_RATING in the subblock is:

EXCELLENT


d. The value of V_CUSTOMER in the main block is:

Womansport


e. The value of V_NAME in the main block is:

V_NAME is not visible in the main block and you would see an error.

f. The value of V_CREDIT_RATING in the main block is:

EXCELLENT




3. Create and execute a PL/SQL block that accepts two numbers through iSQL*Plus substitutionvariables.

a. Use the DEFINE command to provide the two values.

DEFINE p_num1=2 -- example

DEFINE p_num2=4 -- example

b. Pass these two values defined in step a above, to the PL/SQL block through iSQL*Plussubstitution variables. The first number should be divided by the second number and have thesecond number added to the result. The result should be stored in a PL/SQL variable and printedon the screen.

Note: SET VERIFY OFF in the PL/SQL block.

SET ECHO OFF

SET VERIFY OFF

SET SERVEROUTPUT ON

DECLARE

v_num1 NUMBER(9,2) := &p_num1;

v_num2 NUMBER(9,2) := &p_num2;

v_result NUMBER(9,2) ;

BEGIN

v_result := (v_num1/v_num2) + v_num2;

/* Printing the PL/SQL variable */

DBMS_OUTPUT.PUT_LINE (v_result);

END;

/

SET SERVEROUTPUT OFF

SET VERIFY ON

SET ECHO ON



4. Build a PL/SQL block that computes the total compensation for one year.

a. The annual salary and the annual bonus percentage values are defined using the DEFINEcommand.

b. Pass the values defined in the above step to the PL/SQL block through iSQL*Plussubstitution variables. The bonus must be converted from a whole number to a decimal (forexample, 15 to .15). If the salary is null, set it to zero before computing the totalcompensation. Execute the PL/SQL block. Reminder: Use the NVL function to handlenull values.

Note: Total compensation is the sum of the annual salary and the annual bonus.

Method 1: When an iSQL*Plus variable is used:

a. VARIABLE g_total NUMBERDEFINE p_salary=50000DEFINE p_bonus=10

b. SET VERIFY OFF

DECLAREv_salary NUMBER := &p_salary;v_bonus NUMBER := &p_bonus;

BEGIN:g_total := NVL(v_salary, 0) * (1 + NVL(v_bonus, 0) / 100);

END;/PRINT g_totalSET VERIFY ON

Alternate Solution: When a PL/SQL variable is used:

a. DEFINE p_salary=50000DEFINE p_bonus=10

b. SET VERIFY OFFSET SERVEROUTPUT ON

DECLAREv_salary NUMBER := &p_salary;v_bonus NUMBER := &p_bonus;

BEGINdbms_output.put_line(TO_CHAR(NVL(v_salary, 0) *

(1 + NVL(v_bonus, 0) / 100)));END;/SET VERIFY ONSET SERVEROUTPUT OFF


Practice 3 Solutions1. Create a PL/SQL block that selects the maximum department number in the DEPARTMENTS table and

stores it in an iSQL*Plus variable. Print the results to the screen. Save your PL/SQL block in a filenamed p3q1.sql by clicking the Save Script button. Save the script with a .sql extension.

VARIABLE g_max_deptno NUMBER

DECLARE

v_max_deptno NUMBER;

BEGIN

SELECT max(department_id)

INTO v_max_deptno

FROM departments;

:g_max_deptno := v_max_deptno;

END;

/

PRINT g_max_deptno

Alternate Solution:

SET SERVEROUTPUT ON

DECLARE

v_max_deptno NUMBER;

BEGIN

SELECT MAX(department_id) INTO v_max_deptno FROM departments;

dbms_output.put_line(v_max_deptno);

END;

/

2. Modify the PL/SQL block you created in exercise 1 to insert a new department into the DEPARTMENTStable. Save the PL/SQL block in a file named p3q2.sql by clicking the Save Script button. Save thescript with a .sql extension.

a. Use the DEFINE command to provide the department name. Name the new department Education.

SET ECHO OFF

SET VERIFY OFF

DEFINE p_dname = Education

b. Pass the value to the PL/SQL block through a iSQL*Plus substitution variable. Rather than printingthe department number retrieved from exercise 1, add 10 to it and use it as the department number forthe new department.

c. Leave the location number as null for now.



DECLARE

v_max_deptno departments.department_id%TYPE;

BEGIN

SELECT MAX(department_id) + 10

INTO v_max_deptno

FROM departments;

INSERT INTO departments (department_id, department_name,location_id)

VALUES (v_max_deptno, '&p_dname', NULL);

COMMIT;

END;

/SET VERIFY ONSET ECHO ON

d. Execute the PL/SQL block.e. Display the new department that you created.

SELECT *

FROM departments

WHERE department_name = 'Education';

3. Create a PL/SQL block that updates the location ID for the new department that you added in theprevious practice. Save your PL/SQL block in a file named p3q3.sql by clicking the Save Scriptbutton. Save the script with a .sql extension.

a. Use an iSQL*Plus variable for the department ID number that you added in the previous practice.b. Use the DEFINE command to provide the location ID. Name the new location ID 1700.

SET VERIFY OFFDEFINE p_deptno = 280DEFINE p_loc = 1700

c. Pass the value to the PL/SQL block through a iSQL*Plus substitution variable. Test the PL/SQLblock.BEGIN

UPDATE departmentsSET location_id = &p_locWHERE department_id = &p_deptno;COMMIT;

END;/SET VERIFY ON

d. Display the department that you updated.SELECT * FROM departments

WHERE department_id = &p_deptno;



4. Create a PL/SQL block that deletes the department that you created in exercise 2. Save the PL/SQLblock in a file named p3q4.sql by clicking the Save Script button. Save the script with a .sqlextension.

a. Use the DEFINE command to provide the department ID.

SET VERIFY OFF

VARIABLE g_result VARCHAR2(40)

DEFINE p_deptno = 280

b. Pass the value to the PL/SQL block through a iSQL*Plus substitution variable Print to the screen thenumber of rows affected.

c. Test the PL/SQL block.

DECLARE

v_result NUMBER(2);

BEGIN

DELETE

FROM departments


v_result := SQL%ROWCOUNT;

:g_result := (TO_CHAR(v_result) || ' row(s) deleted.');

COMMIT;

END;

/

PRINT g_result

SET VERIFY ON

d. Confirm that the department has been deleted.

SELECT *FROM departmentsWHERE department_id = 280;


Practice 4 Solutions1. Execute the command in the file lab04_1.sql to create the MESSAGES table. Write a PL/SQL

block to insert numbers into the MESSAGES table.

CREATE TABLE messages (results VARCHAR2 (60));

a. Insert the numbers 1 to 10, excluding 6 and 8.

b. Commit before the end of the block.BEGINFOR i IN 1..10 LOOPIF i = 6 or i = 8 THENnull;

ELSEINSERT INTO messages(results)VALUES (i);END IF;COMMIT;

END LOOP;END;/

Note: i is being implicitly converted. A better way to code would be to explicitly convert the NUMBERto VARCHAR2.

c. Select from the MESSAGES table to verify that your PL/SQL block worked.

SELECT *

FROM messages;

2. Create a PL/SQL block that computes the commission amount for a given employee basedon the employee’s salary.

a. Use the DEFINE command to provide the employee ID. Pass the value to the PL/SQL blockthrough a iSQL*Plus substitution variable.

SET SERVEROUTPUT ON

SET VERIFY OFF

DEFINE p_empno = 100

b. If the employee’s salary is less than $5,000, display the bonus amount for the employeeas 10% of the salary.

c. If the employee’s salary is between $5,000 and $10,000, display the bonus amount for theemployee as 15% of the salary.

d. If the employee’s salary exceeds $10,000, display the bonus amount for the employee as20% of the salary.

e. If the employee’s salary is NULL, display the bonus amount for the employee as 0.

f. Test the PL/SQL block for each case using the following test cases, and check eachbonus amount.

Note: Include SET VERIFY OFF in your solution.



DECLARE

v_empno employees.employee_id%TYPE := &p_empno;

v_sal employees.salary%TYPE;

v_bonus_per NUMBER(7,2);

v_bonus NUMBER(7,2);

BEGIN

SELECT salary

INTO v_sal

FROM employees

WHERE employee_id = v_empno;

IF v_sal < 5000 THEN

v_bonus_per := .10;

ELSIF v_sal BETWEEN 5000 and 10000 THEN

v_bonus_per := .15;

ELSIF v_sal > 10000 THEN

v_bonus_per := .20;

ELSE

v_bonus_per := 0;

END IF;

v_bonus := v_sal * v_bonus_per;

DBMS_OUTPUT.PUT_LINE ('The bonus for the employee with employee_id '|| v_empno || ' and salary ' || v_sal || ' is ' || v_bonus);

END;

/



If you have time, complete the following exercises:

3. Create an EMP table that is a replica of the EMPLOYEES table. You can do this by executing thescript lab04_3.sql. Add a new column, STARS, of VARCHAR2 data type and length 50 tothe EMP table for storing asterisk (*).ALTER TABLE empADD stars VARCHAR2(50);

4. Create a PL/SQL block that rewards an employee by appending an asterisk in the STARS columnfor every $1000 of the employee’s salary. Save your PL/SQL block in a file called p4q4.sql byclicking on the Save Script button. Remember to save the script with a .sql extension.

a. Use the DEFINE command to provide the employee ID. Pass the value to the PL/SQL blockthrough a iSQL*Plus substitution variable.

SET VERIFY OFF

DEFINE p_empno = 104

b. Initialize a v_asterisk variable that contains a NULL.

c. Append an asterisk to the string for every $1000 of the salary amount. For example, if theemployee has a salary amount of $8000, the string of asterisks should contain eight asterisks.If the employee has a salary amount of $12500, the string of asterisks should contain 13asterisks.

d. Update the STARS column for the employee with the string of asterisks.

e. Commit.

f. Test the block for the following values:

DEFINE p_empno=104DEFINE p_empno=174DEFINE p_empno=176

Note: SET VERIFY OFF in the PL/SQL block

Employee Number Salary Resulting Bonus

100 24000 4800

149 10500 2100

178 7000 1050



DECLARE

v_empno emp.employee_id%TYPE := TO_NUMBER(&p_empno);

v_asterisk emp.stars%TYPE := NULL;

v_sal emp.salary%TYPE;

BEGIN

SELECT NVL(ROUND(salary/1000), 0)

INTO v_sal

FROM emp


FOR i IN 1..v_sal LOOP

v_asterisk := v_asterisk ||'*';

END LOOP;

UPDATE emp

SET stars = v_asterisk


COMMIT;

END;

/

SET VERIFY ON

g. Display the rows from the EMP table to verify whether your PL/SQL block has executedsuccessfully.

SELECT employee_id,salary, stars

FROM emp

WHERE employee_id IN (104,174,176);



1. Write a PL/SQL block to print information about a given country.

a. Declare a PL/SQL record based on the structure of the COUNTRIES table.

b. Use the DEFINE command to provide the country ID. Pass the value to the PL/SQL blockthrough a iSQL*Plus substitution variable.

SET SERVEROUTPUT ON

SET VERIFY OFF

DEFINE p_countryid = CA

c. Use DBMS_OUTPUT.PUT_LINE to print selected information about the country. A sampleoutput is shown below.

DECLARE

country_record countries%ROWTYPE;

BEGIN

SELECT *

INTO country_record

FROM countries

WHERE country_id = UPPER('&p_countryid');

DBMS_OUTPUT.PUT_LINE ('Country Id: ' ||country_record.country_id || ' Country Name: ' ||country_record.country_name || ' Region: ' ||country_record.region_id);

END;

/

d. Execute and test the PL/SQL block for the countries with the IDs CA, DE, UK, US


Practice 5 Solutions (continued)2. Create a PL/SQL block to retrieve the name of each department from the DEPARTMENTS table and print

each department name on the screen, incorporating an INDEX BY table. Save the code in afile called p5q2.sql by clicking the Save Script button. Save the script with a .sql extension.

a. Declare an INDEX BY table, MY_DEPT_TABLE, to temporarily store the name of thedepartments.

b. Using a loop, retrieve the name of all departments currently in the DEPARTMENTS table and storethem in the INDEX BY table. Use the following table to assign the value for DEPARTMENT_IDbased on the value of the counter used in the loop.

c. Using another loop, retrieve the department names from the PL/SQL table and print them to thescreen, using DBMS_OUTPUT.PUT_LINE.

SET SERVEROUTPUT ONDECLARE

TYPE DEPT_TABLE_TYPE ISTABLE OF departments.department_name%TYPEINDEX BY BINARY_INTEGER;

my_dept_table dept_table_type;v_count NUMBER (2);v_deptno departments.department_id%TYPE;

BEGINSELECT COUNT(*) INTO v_count FROM departments;FOR i IN 1..v_countLOOP

IF i = 1 THENv_deptno := 10;

ELSIF i = 2 THENv_deptno := 20;






END IF;

COUNTER DEPARTMENT_ID

1 102 203 504 605 806 90

7 110



SELECT department_name INTO my_dept_table(i) FROM departmentsWHERE department_id = v_deptno;

END LOOP;FOR i IN 1..v_countLOOPDBMS_OUTPUT.PUT_LINE (my_dept_table(i));

END LOOP;END;/


If you have time, complete the following exercise.

3. Modify the block you created in practice 2 to retrieve all information about each department fromthe DEPARTMENTS table and print the information to the screen, incorporating an INDEX BYtable of records.

a. Declare an INDEX BY table, MY_DEPT_TABLE, to temporarily store the number, name, andlocation of all the departments.

b. Using a loop, retrieve all department information currently in the DEPARTMENTS table andstore it in the PL/SQL table. Use the following table to assign the value for DEPARTMENT_IDbased on the value of the counter used in the loop. Exit the loop when the counter reachesthe value 7.

c. Using another loop, retrieve the department information from the PL/SQL table and print itto the screen, using DBMS_OUTPUT.PUT_LINE.

COUNTER DEPARTMENT_ID

1 102 203 504 605 806 90

7 110


Practice 5 Solutions (continued)SET SERVEROUTPUT ON

DECLARE

TYPE dept_table_type is table of departments%ROWTYPE

INDEX BY BINARY_INTEGER;

my_dept_table dept_table_type;

v_deptno departments.department_id%TYPE;

v_count NUMBER := 7;

BEGIN

FOR i IN 1..v_count

LOOP

IF i = 1 THEN

v_deptno := 10;

ELSIF i = 2 THEN

v_deptno := 20;

ELSIF i = 3 THEN

v_deptno := 50;

ELSIF i = 4 THEN

v_deptno := 60;

ELSIF i = 5 THEN

v_deptno := 80;

ELSIF i = 6 THEN

v_deptno := 90;

ELSIF i = 7 THEN

v_deptno := 110;

END IF;

SELECT *

INTO my_dept_table(i)

FROM departments

WHERE department_id = v_deptno;

END LOOP;

FOR i IN 1..v_count

LOOP

DBMS_OUTPUT.PUT_LINE ('Department Number: ' ||my_dept_table(i).department_id

|| ' Department Name: ' || my_dept_table(i).department_name

|| ' Manager Id: '|| my_dept_table(i).manager_id

|| ' Location Id: ' || my_dept_table(i).location_id);

END LOOP;

END;

/


Practice 6 Solutions1. Run the command in the script lab06_1.sql to create a new table for storing the salaries of the

employees.

CREATE TABLE top_dogs

(salary NUMBER(8,2));

2. Create a PL/SQL block that determines the top employees with respect to salaries.

a. Accept a number n from the user where n represents the number of top n earners from theEMPLOYEES table. For example, to view the top five earners, enter 5.

Note: Use the DEFINE command to provide the value for n. Pass the value to the PL/SQLblock through a iSQL*Plus substitution variable.

DELETE FROM top_dogs;

DEFINE p_num = 5

b. In a loop use the iSQL*Plus substitution parameter created in step 1 and gather the salaries ofthe top n people from the EMPLOYEES table. There should be no duplication in the salaries. Iftwo employees earn the same salary, the salary should be picked up only once.

c. Store the salaries in the TOP_DOGS table.

d. Test a variety of special cases, such as n = 0 or where n is greater than the number of employeesin the EMPLOYEES table. Empty the TOP_DOGS table after each test. The output shownrepresents the five highest salaries in the EMPLOYEES table.

DECLARE

v_num NUMBER(3) := &p_num;


CURSOR emp_cursor IS

SELECT distinct salary

FROM employees

ORDER BY salary DESC;

BEGIN

OPEN emp_cursor;

FETCH emp_cursor INTO v_sal;

WHILE emp_cursor%ROWCOUNT <= v_num AND emp_cursor%FOUND LOOP

INSERT INTO top_dogs (salary)

VALUES (v_sal);

FETCH emp_cursor INTO v_sal;

END LOOP;

CLOSE emp_cursor;

COMMIT;

END;

/

SELECT * FROM top_dogs:



3. Create a PL/SQL block that does the following:

a. Use the DEFINE command to provide the department ID. Pass the value to the PL/SQL blockthrough a iSQL*Plus substitution variable.

SET SERVEROUTPUT ON

SET ECHO OFF

DEFINE p_dept_no = 10

b. In a PL/SQL block, retrieve the last name, salary and MANAGER ID of the employees workingin that department.

c. If the salary of the employee is less than 5000 and if the manager ID is either 101 or 124,display the message <<last_name>> Due for a raise. Otherwise, display amessage <<last_name>> Not due for a raise.

Note: SET ECHO OFF to avoid displaying the PL/SQL code every time you execute the script

d. Test the PL/SQL block for the following cases:

Department ID Message

10 Whalen Due for a raise

20 Hartstein Not Due for a raiseFay Not Due for a raise

50 Weiss Not Due for a raiseFripp Due for a raiseKaufling Due for a raiseVollman Due for a raiseMourgas Due for a raise. . .. . .

80 Russel Not Due for a raisePartners Not Due for a raiseErrazuriz Not Due for a raiseCambrault Not Due for a raise. . .. . .



DECLARE

v_deptno NUMBER(4) := &p_dept_no;

v_ename employees.last_name%TYPE;


v_manager employees.manager_id%TYPE;

CURSOR emp_cursor IS

SELECT last_name, salary,manager_id

FROM employees

WHERE department_id = v_deptno;

BEGIN

OPEN emp_cursor;

FETCH emp_cursor INTO v_ename, v_sal,v_manager;

WHILE emp_cursor%FOUND LOOP

IF v_sal < 5000 AND (v_manager = 101 OR v_manager = 124) THEN

DBMS_OUTPUT.PUT_LINE (v_ename || ' Due for a raise');

ELSE

DBMS_OUTPUT.PUT_LINE (v_ename || ' Not Due for a raise');

END IF;

FETCH emp_cursor INTO v_ename, v_sal,v_manager;

END LOOP;

CLOSE emp_cursor;

END;

/




1. In a loop, use a cursor to retrieve the department number and the department name from theDEPARTMENTS table for those departments whose DEPARTMENT_ID is less than 100. Pass thedepartment number to another cursor to retrieve from the EMPLOYEES table the details of employeelast name, job, hire date, and salary of those employees whose EMPLOYEE_ID is less than 120 andwho work in that department.

SET SERVEROUTPUT ON

DECLARE

CURSOR dept_cursor IS

SELECT department_id,department_name

FROM departments

WHERE department_id < 100


CURSOR emp_cursor(v_deptno NUMBER) IS

SELECT last_name,job_id,hire_date,salary

FROM employees

WHERE department_id = v_deptno

AND employee_id < 120;

v_current_deptno departments.department_id%TYPE;

v_current_dname departments.department_name%TYPE;


v_job employees.job_id%TYPE;

v_hiredate employees.hire_date%TYPE;


v_line varchar2(100);

BEGIN

v_line := '';

OPEN dept_cursor;

LOOP

FETCH dept_cursor INTO v_current_deptno,v_current_dname;

EXIT WHEN dept_cursor%NOTFOUND;

DBMS_OUTPUT.PUT_LINE ('Department Number : ' ||v_current_deptno || ' Department Name : ' || v_current_dname);



DBMS_OUTPUT.PUT_LINE(v_line);

IF emp_cursor%ISOPEN THEN

CLOSE emp_cursor;

END IF;

OPEN emp_cursor (v_current_deptno);

LOOP

FETCH emp_cursor INTOv_ename,v_job,v_hiredate,v_sal;

EXIT WHEN emp_cursor%NOTFOUND;

DBMS_OUTPUT.PUT_LINE (v_ename || ' ' || v_job || ' '|| v_hiredate || ' ' || v_sal);

END LOOP;


CLOSE emp_cursor;

END IF;

DBMS_OUTPUT.PUT_LINE(v_line);

END LOOP;


CLOSE emp_cursor;

END IF;

CLOSE dept_cursor;

END;

/


Alternative Solution:

SET SERVEROUTPUT ON

DECLARE

CURSOR DEPT_CUR IS

SELECT DEPARTMENT_ID DEPTNO, DEPARTMENT_NAME DNAME

FROM DEPARTMENTS

WHERE DEPARTMENT_ID < 100;

CURSOR EMP_CUR (P_DEPTNO NUMBER) IS

SELECT * FROM EMPLOYEES

WHERE DEPARTMENT_ID = P_DEPTNO AND EMPLOYEE_ID < 120;



BEGIN

FOR DEPT_REC IN DEPT_CUR LOOP

DBMS_OUTPUT.PUT_LINE

('DEPARTMENT NUMBER: ' || DEPT_REC.DEPTNO ||'

DEPARTMENT NAME: ' || DEPT_REC.DNAME);

FOR EMP_REC IN EMP_CUR(DEPT_REC.DEPTNO) LOOP

DBMS_OUTPUT.PUT_LINE

(EMP_REC.LAST_NAME ||' '||EMP_REC.JOB_ID||'

'||EMP_REC.HIRE_DATE||' '||EMP_REC.SALARY);

END LOOP;

DBMS_OUTPUT.PUT_LINE(CHR(10));

END LOOP;

END;

/


Practice 7 Solutions (continued)2. Modify the code in sol04_4.sql to incorporate a cursor using the FOR UPDATE and WHERE

CURRENT OF functionality in cursor processing.

a. Define the host variables.

SET VERIFY OFFDEFINE p_empno = 104

b. Execute the modified PL/SQL blockDECLARE

v_empno emp.employee_id%TYPE := &p_empno;v_asterisk emp.stars%TYPE := NULL;CURSOR emp_cursor ISSELECT employee_id, NVL(ROUND(salary/1000), 0) salFROM empWHERE employee_id = v_empnoFOR UPDATE;

BEGINFOR emp_record IN emp_cursor LOOPFOR i IN 1..emp_record.sal LOOPv_asterisk := v_asterisk ||'*';DBMS_OUTPUT.PUT_LINE(v_asterisk);

END LOOP;UPDATE empSET stars = v_asteriskWHERE CURRENT OF emp_cursor;v_asterisk := NULL;

END LOOP;COMMIT;

END;/SET VERIFY ON

c. Execute the following command to check if your PL/SQL block has worked successfully:

SELECT employee_id,salary,starsFROM EMPWHERE employee_id IN (176,174,104);



1. Write a PL/SQL block to select the name of the employee with a given salary value.

a. Use the DEFINE command to provide the salary.

SET VERIFY OFF

DEFINE p_sal = 6000

b. Pass the value to the PL/SQL block through a iSQL*Plus substitution variable. If the salaryentered returns more than one row, handle the exception with an appropriate exceptionhandler and insert into the MESSAGES table the message “More than one employee with asalary of <salary>.”

c. If the salary entered does not return any rows, handle the exception with an appropriateexception handler and insert into the MESSAGES table the message “No employee with asalary of <salary>.”

d. If the salary entered returns only one row, insert into the MESSAGES table the employee’sname and the salary amount.

e. Handle any other exception with an appropriate exception handler and insert into theMESSAGES table the message “Some other error occurred.”

f. Test the block for a variety of test cases. Display the rows from the MESSAGES table to checkwhether the PL/SQL block has executed successfully

DECLARE


v_sal employees.salary%TYPE := &p_sal;

BEGIN

SELECT last_name

INTO v_ename

FROM employees

WHERE salary = v_sal;

INSERT INTO messages (results)

VALUES (v_ename || ' - ' || v_sal);

EXCEPTION

WHEN no_data_found THEN


VALUES ('No employee with a salary of '|| TO_CHAR(v_sal));

WHEN too_many_rows THEN


VALUES ('More than one employee with a salary of '||

TO_CHAR(v_sal));

WHEN others THEN


VALUES ('Some other error occurred.');

END;

/

SET VERIFY ON


Practice 8 Solutions (continued)2. Modify the code in p3q3.sql to add an exception handler.

a. Use the DEFINE command to provide the department ID and department location. Pass thevalues to the PL/SQL block through a iSQL*Plus substitution variables.

SET VERIFY OFF

VARIABLE g_message VARCHAR2(100)

DEFINE p_deptno = 200

DEFINE p_loc = 1400

b. Write an exception handler for the error to pass a message to the user that the specifieddepartment does not exist. Use a bind variable to pass the message to the user.

c. Execute the PL/SQL block by entering a department that does not exist.

DECLARE

e_invalid_dept EXCEPTION;

v_deptno departments.department_id%TYPE := &p_deptno;

BEGIN

UPDATE departments

SET location_id = &p_loc


COMMIT;

IF SQL%NOTFOUND THEN

raise e_invalid_dept;

END IF;

EXCEPTION

WHEN e_invalid_dept THEN

:g_message := 'Department '|| TO_CHAR(v_deptno) ||' is aninvalid department';

END;

/

SET VERIFY ON

PRINT g_message



3. Write a PL/SQL block that prints the number of employees who earn plus or minus $100 of the salaryvalue set for an iSQL*Plus substitution variable. Use the DEFINE command to provide the salaryvalue. Pass the value to the PL/SQL block through a iSQL*Plus substitution variable.

a. If there is no employee within that salary range, print a message to the user indicatingthat is the case. Use an exception for this case.

VARIABLE g_message VARCHAR2(100)

SET VERIFY OFF

DEFINE p_sal = 7000

b. If there are one or more employees within that range, the message should indicate how manyemployees are in that salary range.

c. Handle any other exception with an appropriate exception handler. The message should indicatethat some other error occurred.

DECLARE

v_sal employees.salary%TYPE := &p_sal;

v_low_sal employees.salary%TYPE := v_sal - 100;

v_high_sal employees.salary%TYPE := v_sal + 100;

v_no_emp NUMBER(7);

e_no_emp_returned EXCEPTION;

e_more_than_one_emp EXCEPTION;

BEGIN

SELECT count(last_name)

INTO _no_emp

FROM employees

WHERE salary between v_low_sal and v_high_sal;

IF v_no_emp = 0 THEN

RAISE e_no_emp_returned;

ELSIF v_no_emp > 0 THEN

RAISE e_more_than_one_emp;

END IF;


Practice 8 Solutions (continued)EXCEPTION

WHEN e_no_emp_returned THEN

:g_message := 'There is no employee salary between '||

TO_CHAR(v_low_sal) || ' and '||

TO_CHAR(v_high_sal);

WHEN e_more_than_one_emp THEN

:g_message := 'There is/are '|| TO_CHAR(v_no_emp) ||

' employee(s) with a salary between '||

TO_CHAR(v_low_sal) || ' and '||

TO_CHAR(v_high_sal);

WHEN others THEN

:g_message := 'Some other error occurred.';

END;

/

SET VERIFY ON

PRINT g_message


Practice 9 SolutionsNote: Save your subprograms as .sql files, using the Save Script button.

Remember to set the SERVEROUTPUT ON if you set it off previously.

1. Create and invoke the ADD_JOB procedure and consider the results.

a. Create a procedure called ADD_JOB to insert a new job into the JOBS table. Provide the ID andtitle of the job, using two parameters.

CREATE OR REPLACE PROCEDURE add_job

(p_jobid IN jobs.job_id%TYPE,

p_jobtitle IN jobs.job_title%TYPE)

IS

BEGIN

INSERT INTO jobs (job_id, job_title)

VALUES (p_jobid, p_jobtitle);

COMMIT;

END add_job;

b. Compile the code, and invoke the procedure with IT_DBA as job ID and DatabaseAdministrator as job title. Query the JOBS table to view the results.

In iSQL*Plus, load and run the script file created in question 1a above.

EXECUTE add_job ('IT_DBA', 'Database Administrator')

SELECT * FROM jobs WHERE job_id = 'IT_DBA';

c. Invoke your procedure again, passing a job ID of ST_MAN and a job title of Stock Manager.What happens and why?

EXECUTE add_job ('ST_MAN', 'Stock Manager')

There is a primary key integrity constraint on the JOB_ID column.


Practice 9 Solutions (continued)2. Create a procedure called UPD_JOB to modify a job in the JOBS table.

a. Create a procedure called UPD_JOB to update the job title. Provide the job ID and a new title,using two parameters. Include the necessary exception handling if no update occurs.

CREATE OR REPLACE PROCEDURE upd_job


p_jobtitle IN jobs.job_title%TYPE)

IS

BEGIN

UPDATE jobs

SET job_title = p_jobtitle

WHERE job_id = p_jobid;


RAISE_APPLICATION_ERROR(-20202,'No job updated.');

END IF;

END upd_job;

/

b. Compile the code; invoke the procedure to change the job title of the job ID IT_DBA to DataAdministrator. Query the JOBS table to view the results. Also check the exceptionhandling by trying to update a job that does not exist (you can use job ID IT_WEB and job titleWeb Master).

In iSQL*Plus, load and run the script file created in the above question.

EXECUTE upd_job ('IT_DBA', 'Data Administrator')


EXECUTE upd_job ('IT_WEB', 'Web Master')


Practice 9 Solutions (continued)3. Create a procedure called DEL_JOB to delete a job from the JOBS table.

a. Create a procedure called DEL_JOB to delete a job from the JOBS table. Include thenecessary exception handling if no job is deleted.

CREATE OR REPLACE PROCEDURE del_job

(p_jobid IN jobs.job_id%TYPE)

IS

BEGIN

DELETE FROM jobs



RAISE_APPLICATION_ERROR(-20203,'No jobs deleted.');

END IF;

END DEL_JOB;

/

b. Compile the code; invoke the procedure using job ID IT_DBA. Query the JOBS table toview the results.


EXECUTE del_job ('IT_DBA')


Also, check the exception handling by trying to delete a job that does not exist (use job IDIT_WEB). You should get the message you used in the exception-handling section of the procedureas output.

EXECUTE del_job ('IT_WEB')


Practice 9 Solutions (continued)4. Create a procedure called QUERY_EMP to query the EMPLOYEES table, retrieving the salary and job

ID for an employee when provided with the employee ID.

a. Create a procedure that returns a value from the SALARY and JOB_ID columns for aspecified employee ID.

Use host variables for the two OUT parameters salary and job ID.

CREATE OR REPLACE PROCEDURE query_emp

(p_empid IN employees.employee_id%TYPE,

p_sal OUT employees.salary%TYPE,

p_job OUT employees.job_id%TYPE)

IS

BEGIN

SELECT salary, job_id

INTO p_sal, p_job

FROM employees

WHERE employee_id = p_empid;

END query_emp;

/

b. Compile the code, invoke the procedure to display the salary and job ID for employee ID 120.


VARIABLE g_sal NUMBER

VARIABLE g_job VARCHAR2(15)

EXECUTE query_emp (120, :g_sal, :g_job)

PRINT g_sal

PRINT g_job



c. Invoke the procedure again, passing an EMPLOYEE_ID of 300. What happens and why?

EXECUTE query_emp (300, :g_sal, :g_job)

There is no employee in the EMPLOYEES table with an EMPLOYEE_ID of 300. The SELECTstatement retrieved no data from the database, resulting in a fatal PL/SQL error,NO_DATA_FOUND.


Practice 10 Solutions1. Create and invoke the Q_JOB function to return a job title.

a. Create a function called Q_JOB to return a job title to a host variable.

CREATE OR REPLACE FUNCTION q_job


RETURN VARCHAR2

IS

v_jobtitle jobs.job_title%TYPE;

BEGIN

SELECT job_title

INTO v_jobtitle

FROM jobs


RETURN (v_jobtitle);

END q_job;

/

b. Compile the code; create a host variable G_TITLE and invoke the function with job ID SA_REP.Query the host variable to view the result.


VARIABLE g_title VARCHAR2(30)

EXECUTE :g_title := q_job ('SA_REP')

PRINT g_title


Practice 10 Solutions (continued)2. Create a function called ANNUAL_COMP to return the annual salary by accepting two parameters: an

employee’s monthly salary and commission. The function should address NULL values.

a. Create and invoke the function ANNUAL_COMP, passing in values for monthly salary andcommission. Either or both values passed can be NULL, but the function should still return anannual salary, which is not NULL. The annual salary is defined by the basic formula:

(salary*12) + (commission_pct*salary*12)

CREATE OR REPLACE FUNCTION annual_comp

(p_sal IN employees.salary%TYPE,

p_comm IN employees.commission_pct%TYPE)

RETURN NUMBER

IS

BEGIN

RETURN (NVL(p_sal,0) * 12 + (NVL(p_comm,0)* p_sal * 12));

END annual_comp;

/

b. Use the function in a SELECT statement against the EMPLOYEES table for department 80.

SELECT employee_id, last_name, annual_comp(salary,commission_pct)"Annual Compensation"

FROM employees

WHERE department_id=80;

…


Practice 10 Solutions (continued)3. Create a procedure, NEW_EMP, to insert a new employee into the EMPLOYEES table. The procedure

should contain a call to the VALID_DEPTID function to check whether the department ID specifiedfor the new employee exists in the DEPARTMENTS table.

a. Create a function VALID_DEPTID to validate a specified department ID. The function shouldreturn a BOOLEAN value.

CREATE OR REPLACE FUNCTION valid_deptid

(p_deptid IN departments.department_id%TYPE)

RETURN BOOLEAN

IS

v_dummy VARCHAR2(1);

BEGIN

SELECT 'x'

INTO v_dummy

FROM departments

WHERE department_id = p_deptid;

RETURN (TRUE);

EXCEPTION

WHEN NO_DATA_FOUND THEN

RETURN (FALSE);

END valid_deptid;

/


Practice 10 Solutions (continued)b. Create the procedure NEW_EMP to add an employee to the EMPLOYEES table. A new row should

be added to EMPLOYEES table if the function returns TRUE. If the function returns FALSE, theprocedure should alert the user with an appropriate message.

Define DEFAULT values for most parameters. The default commission is 0, the default salary is1000, the default department ID is 30, the default job is SA_REP and the default manager ID is145. For the employee’s ID, use the sequence EMPLOYEES _SEQ. Provide the last name, firstname and e-mail address of the employee.

CREATE OR REPLACE PROCEDURE new_emp

(p_lname employees.last_name%TYPE,

p_fname employees.first_name%TYPE,

p_email employees.email%TYPE,

p_job employees.job_id%TYPE DEFAULT 'SA_REP',

p_mgr employees.manager_id%TYPE DEFAULT 145,

p_sal employees.salary%TYPE DEFAULT 1000,

p_comm employees.commission_pct%TYPE DEFAULT 0,

p_deptid employees.department_id%TYPE DEFAULT 30)

IS

BEGIN

IF valid_deptid(p_deptid) THEN

INSERT INTO employees(employee_id, last_name, first_name,email, job_id, manager_id, hire_date,salary, commission_pct, department_id)

VALUES (employees_seq.NEXTVAL, p_lname, p_fname, p_email,p_job, p_mgr, TRUNC (SYSDATE, 'DD'), p_sal,p_comm, p_deptid);

ELSE

RAISE_APPLICATION_ERROR (-20204,

'Invalid department ID. Try again.');

END IF;

END new_emp;

/


Practice 10 Solutions (continued)c. Test your NEW_EMP procedure by adding a new employee named Jane Harris to department 15.

Allow all other parameters to default. What was the result?

EXECUTE new_emp(p_lname=>'Harris', p_fname=>'Jane',p_email=>'JAHARRIS', p_deptid => 15)

d. Test your NEW_EMP procedure by adding a new employee named Joe Harris to department 80.Allow all other parameters to default. What was the result?

EXECUTE new_emp(p_lname=>'Harris',p_fname=>'Joe',p_email=>'JOHARRIS',p_deptno => 80)


Practice 11 SolutionsSuppose you have lost the code for the NEW_EMP procedure and the VALID_DEPTID function thatyou created in lesson 10. (If you did not complete the practices in lesson 10, you can run the solutionscripts to create the procedure and function.)

Create an iSQL*Plus spool file to query the appropriate data dictionary view to regenerate the code.

Hint:

SET -- options ON|OFF

SELECT -- statement(s) to extract the code

SET -- reset options ON|OFF

To spool the output of the file to a .sql file from iSQL*Plus, select the Save option for theOutput and execute the code.

SET ECHO OFF HEADING OFF FEEDBACK OFF VERIFY OFF

COLUMN LINE NOPRINT

SET PAGESIZE 0

SELECT 'CREATE OR REPLACE ', 0 line

FROM DUAL

UNION

SELECT text, line

FROM USER_SOURCE

WHERE name IN ('NEW_EMP', 'VALID_DEPTNO')

ORDER BY line;

SELECT '/'

FROM DUAL;

SET PAGESIZE 24

COLUMN LINE CLEAR

SET FEEDBACK ON VERIFY ON HEADING ON


Practice 12 Solutions1. Create a package specification and body called JOB_PACK. (You can save the package body and

specification in two separate files.) This package contains your ADD_JOB, UPD_JOB, andDEL_JOB procedures, as well as your Q_JOB function.

Note: Use the code in your previously saved script files when creating the package.

a. Make all the constructs public.

Note: Consider whether you still need the stand-alone procedures and functions you just packaged.

CREATE OR REPLACE PACKAGE job_pack IS

PROCEDURE add_job


p_jobtitle IN jobs.job_title%TYPE);

PROCEDURE upd_job


p_jobtitle IN jobs.job_title%TYPE);

PROCEDURE del_job

(p_jobid IN jobs.job_id%TYPE);

FUNCTION q_job


RETURN VARCHAR2;

END job_pack;

/



CREATE OR REPLACE PACKAGE BODY job_pack ISPROCEDURE add_job(p_jobid IN jobs.job_id%TYPE,p_jobtitle IN jobs.job_title%TYPE)

ISBEGININSERT INTO jobs (job_id, job_title)VALUES (p_jobid, p_jobtitle);

END add_job;PROCEDURE upd_job(p_jobid IN jobs.job_id%TYPE,p_jobtitle IN jobs.job_title%TYPE)

ISBEGINUPDATE jobsSET job_title = p_jobtitleWHERE job_id = p_jobid;IF SQL%NOTFOUND THENRAISE_APPLICATION_ERROR(-20202,'No job updated.');

END IF;END upd_job;

PROCEDURE del_job(p_jobid IN jobs.job_id%TYPE)

ISBEGINDELETE FROM jobsWHERE job_id = p_jobid;IF SQL%NOTFOUND THENRAISE_APPLICATION_ERROR (-20203,'No job deleted.');

END IF;END del_job;

FUNCTION q_job(p_jobid IN jobs.job_id%TYPE)RETURN VARCHAR2

ISv_jobtitle jobs.job_title%TYPE;

BEGINSELECT job_titleINTO v_jobtitleFROM jobsWHERE job_id = p_jobid;RETURN (v_jobtitle);

END q_job;END job_pack;/



b. Invoke your ADD_JOB procedure by passing values IT_SYSAN and SYSTEMS ANALYST asparameters.

EXECUTE job_pack.add_job('IT_SYSAN', 'Systems Analyst')

c. Query the JOBS table to see the result.

SELECT * FROM jobs

WHERE job_id = 'IT_SYSAN';

2. Create and invoke a package that contains private and public constructs.

a. Create a package specification and package body called EMP_PACK that contains your NEW_EMPprocedure as a public construct, and your VALID_DEPTID function as a private construct. (Youcan save the specification and body into separate files.)

CREATE OR REPLACE PACKAGE emp_pack ISPROCEDURE new_emp

(p_lname employees.last_name%TYPE,

p_fname employees.first_name%TYPE,

p_email employees.email%TYPE,

p_job employees.job_id%TYPE DEFAULT 'SA_REP',

p_mgr employees.manager_id%TYPE DEFAULT 145,

p_sal employees.salary%TYPE DEFAULT 1000,

p_comm employees.commission_pct%TYPE DEFAULT 0,p_deptid employees.department_id%TYPE DEFAULT 80);

END emp_pack;

/


Practice 12 Solutions (continued)CREATE OR REPLACE PACKAGE BODY emp_pack IS

FUNCTION valid_deptid

(p_deptid IN departments.department_id%TYPE)RETURN BOOLEAN

IS

v_dummy VARCHAR2(1);

BEGIN

SELECT 'x'

INTO v_dummy

FROM departments


RETURN (TRUE);

EXCEPTION


RETURN(FALSE);END valid_deptid;PROCEDURE new_emp(p_lname employees.last_name%TYPE,p_fname employees.first_name%TYPE,p_email employees.email%TYPE,p_job employees.job_id%TYPE DEFAULT 'SA_REP',p_mgr employees.manager_id%TYPE DEFAULT 145,p_sal employees.salary%TYPE DEFAULT 1000,p_comm employees.commission_pct%TYPE DEFAULT 0,p_deptid employees.department_id%TYPE DEFAULT 80)

ISBEGINIF valid_deptid(p_deptid) THENINSERT INTO employees (employee_id, last_name, first_name,email, job_id, manager_id, hire_date, salary, commission_pct,department_id)

VALUES (employees_seq.NEXTVAL, p_lname, p_fname, p_email,p_job, p_mgr, TRUNC (SYSDATE, 'DD'), p_sal, p_comm,p_deptid);

ELSERAISE_APPLICATION_ERROR (-20205,'Invalid department number. Try again.');

END IF;END new_emp;

END emp_pack;/



b. Invoke the NEW_EMP procedure, using 15 as a department number. As the department ID 15does not exist in the DEPARTMENTS table, you should get an error message as specified in theexception handler of your procedure.

EXECUTE emp_pack.new_emp(p_lname=>'Harris',p_fname=>'Jane',p_email=>'JAHARRIS', p_deptid => 15)

c. Invoke the NEW_EMP procedure, using an exising department ID 80.

EXECUTE emp_pack.new_emp(p_lname =>'Smith', p_fname=>'David',p_email=>'DASMITH', p_deptid=>80)

If you have time:3. a. Create a package called CHK_PACK that contains the procedures CHK_HIREDATE and

CHK_DEPT_MGR. Make both constructs public. (You can save the specification and body intoseparate files.)The procedure CHK_HIREDATE checks whether an employee’s hire date is within the followingrange: [SYSDATE - 50 years, SYSDATE + 3 months].

Note:• If the date is invalid, you should raise an application error with an appropriate message

indicating why the date value is not acceptable.• Make sure the time component in the date value is ignored.• Use a constant to refer to the 50 years boundary.• A null value for the hire date should be treated as an invalid hire date.

The procedure CHK_DEPT_MGR checks the department and manager combination for a givenemployee. The CHK_DEPT_MGR procedure accepts an employee ID and a manager ID. Theprocedure checks that the manager and employee work in the same department. The procedure alsochecks that the job title of the manager ID provided is MANAGER.

Note: If the department ID and manager combination is invalid, you should raise an application errorwith an appropriate message.CREATE OR REPLACE PACKAGE chk_pack ISPROCEDURE chk_hiredate

(p_date in employees.hire_date%type);PROCEDURE chk_dept_mgr

(p_empid in employees.employee_id%type,p_mgr in employees.manager_id%type);

END chk_pack;/



CREATE OR REPLACE PACKAGE BODY chk_pack IS

PROCEDURE chk_hiredate(p_date in employees.hire_date%TYPE)ISv_low date := ADD_MONTHS (SYSDATE, - (50 * 12));v_high date := ADD_MONTHS (SYSDATE, 3);

BEGINIF TRUNC(p_date) NOT BETWEEN v_low AND v_high

OR p_date IS NULL THENRAISE_APPLICATION_ERROR(-20200,'Not a valid hiredate');

END IF;END chk_hiredate;

PROCEDURE chk_dept_mgr(p_empid in employees.employee_id%TYPE,p_mgr in employees.manager_id%TYPE)

ISv_empnr employees.employee_id%TYPE;v_deptid employees.department_id%TYPE;

BEGINBEGINSELECT department_idINTO v_deptidFROM employeesWHERE employee_id = p_empid;

EXCEPTIONWHEN NO_DATA_FOUNDTHEN RAISE_APPLICATION_ERROR(-20000, 'Not a valid emp id');

END;BEGINSELECT employee_id /*check valid combination

deptno/mgr for given employee */INTO v_empnrFROM employeesWHERE department_id = v_deptidAND employee_id = p_mgrAND job_id like '%MAN';

EXCEPTIONWHEN NO_DATA_FOUNDTHEN RAISE_APPLICATION_ERROR (-20000,

'Not a valid manager for this department');END;

END chk_dept_mgr;END chk_pack;/


Practice 12 Solutions (continued)b. Test the CHK_HIREDATE procedure with the following command:

EXECUTE chk_pack.chk_hiredate('01-JAN-47')


c. Test the CHK_HIREDATE procedure with the following command:

EXECUTE chk_pack.chk_hiredate(NULL)


d. Test the CHK_DEPT_MGR procedure with the following command:

EXECUTE chk_pack.chk_dept_mgr(117, 100)



Practice 13 Solutions1. Create a package called OVER_LOAD. Create two functions in this package; name each function

PRINT_IT. The function accepts a date or a character string and prints a date or a number, dependingon how the function is invoked.

Note:

• To print the date value, use DD-MON-YY as the input format, and FmMonth,dd yyyy as theoutput format. Make sure you handle invalid input.

• To print out the number, use 999,999.00 as the input format.

The package specification:

CREATE OR REPLACE PACKAGE over_load IS

FUNCTION print_it(p_arg IN DATE)

RETURN VARCHAR2;

FUNCTION print_it(p_arg IN VARCHAR2)

RETURN NUMBER;

END over_load;

/

The package body:CREATE OR REPLACE PACKAGE BODY over_load

IS

FUNCTION print_it(p_arg IN DATE)

RETURN VARCHAR2

IS

BEGIN

RETURN to_char(p_arg, 'FmMonth,dd yyyy');

END print_it;

FUNCTION print_it(p_arg IN VARCHAR2)

RETURN NUMBER

IS

BEGIN

RETURN TO_NUMBER(p_arg, '999,999.00');

-- or use the NLS characters for grands and decimals

-- RETURN TO_NUMBER(p_arg, '999G999D00');

END print_it;

END over_load;

/


Practice 13 Solutions (continued)a. Test the first version of PRINT_IT with the following set of commands:

VARIABLE display_date VARCHAR2(20)

EXECUTE :display_date := over_load.print_it(TO_DATE('08-MAR-01'))

PRINT display_date

b. Test the second version of PRINT_IT with the following set of commands:

VARIABLE g_emp_sal number

EXECUTE :g_emp_sal := over_load.print_it('33,600')

PRINT g_emp_sal

2. Create a new package, called CHECK_PACK, to implement a new business rule.

a. Create a procedure called CHK_DEPT_JOB to verify whether a given combination of departmentID and job is a valid one. In this case valid means that it must be a combination that currentlyexists in the EMPLOYEES table.

Note:

• Use a PL/SQL table to store the valid department and job combination.

• The PL/SQL table needs to be populated only once.

• Raise an application error with an appropriate message if the combination is not valid.

CREATE OR REPLACE PACKAGE check_pack IS

PROCEDURE chk_dept_job

(p_deptid IN employees.department_id%TYPE,

p_job IN employees.job_id%TYPE);

END check_pack;

/



CREATE OR REPLACE PACKAGE BODY check_pack

IS

i NUMBER := 0;

CURSOR emp_cur IS

SELECT department_id, job_id

FROM employees;

TYPE emp_table_type IS TABLE OF emp_cur%ROWTYPE

INDEX BY BINARY_INTEGER;

deptid_job emp_table_type;

PROCEDURE chk_dept_job

(p_deptid in employees.department_id%TYPE,

p_job in employees.job_id%TYPE)

IS

BEGIN

FOR k IN deptid_job.FIRST .. deptid_job.LAST LOOP

IF p_deptid = deptid_job(k).department_id

AND p_job = deptid_job(k).job_id THEN

RETURN;

END IF;

END LOOP;

RAISE_APPLICATION_ERROR

(-20500, 'Not a valid job for this dept');

END chk_dept_job;

BEGIN -- one-time-only-procedure

FOR emp_rec IN emp_cur LOOP

deptid_job(i) := emp_rec;

i := i + 1;

END LOOP;

END check_pack;

/


Practice 13 Solutions (continued)b. Test your CHK_DEPT_JOB package procedure by executing the following command:


What happens?

c. Test your CHK_DEPT_JOB package procedure by executing the following command:




Practice 14 Solutions1a. Create a procedure DROP_TABLE that drops the table specified in the input parameter. Use

the procedures and functions from the supplied DBMS_SQL package.

CREATE OR REPLACE PROCEDURE drop_table

(p_table_name IN VARCHAR2)

IS

dyn_cur NUMBER;

dyn_err VARCHAR2(255);

BEGIN

dyn_cur := DBMS_SQL.OPEN_CURSOR;

DBMS_SQL.PARSE(dyn_cur, 'DROP TABLE '||

p_table_name, DBMS_SQL.NATIVE);

DBMS_SQL.CLOSE_CURSOR(dyn_cur);

EXCEPTION

WHEN OTHERS THEN dyn_err := SQLERRM;

DBMS_SQL.CLOSE_CURSOR(dyn_cur);

RAISE_APPLICATION_ERROR(-20600, dyn_err);

END drop_table;

/

b. To test the DROP_TABLE procedure, first create a new table called EMP_DUP as a copy ofthe EMPLOYEES table.

CREATE TABLE emp_dup AS

SELECT * FROM employees;

c. Execute the DROP_TABLE procedure to drop the EMP_DUP table.

EXECUTE drop_table('emp_dup')

SELECT * FROM emp_dup;


Practice 14 Solutions (continued)2a. Create another procedure called DROP_TABLE2 that drops the table specified in the input

parameter. Use the EXECUTE IMMEDIATE statement.

CREATE PROCEDURE DROP_TABLE2

(p_table_name IN VARCHAR2)

IS

BEGIN

EXECUTE IMMEDIATE 'DROP TABLE ' || p_table_name;

END;

/

b. Repeat the test outlined in steps 1b and 1c.

CREATE TABLE emp_dup AS


EXECUTE drop_table2('emp_dup')

SELECT * FROM emp_dup;


Practice 14 Solutions (continued)3a. Create a procedure called ANALYZE_OBJECT that analyzes the given object that you specified in

the input parameters. Use the DBMS_DDL package, and use the COMPUTE method.

CREATE OR REPLACE procedure analyze_object

(p_obj_type IN VARCHAR2,

p_obj_name IN VARCHAR2)

IS

BEGIN

DBMS_DDL.ANALYZE_OBJECT(

p_obj_type,

USER,

UPPER(p_obj_name),

'COMPUTE');

END;

/

b. Test the procedure using the table EMPLOYEES.

Confirm that the ANALYZE_OBJECT procedure has run by querying the LAST_ANALYZEDcolumn in the USER_TABLES data dictionary view.

EXECUTE ANALYZE_OBJECT ('TABLE', 'EMPLOYEES')

SELECT LAST_ANALYZED FROM USER_TABLES

WHERE TABLE_NAME = 'EMPLOYEES';



If you have time:

4a. Schedule ANALYZE_OBJECT by using DBMS_JOB. Analyze the DEPARTMENTS table, andschedule the job to run in five minutes time from now. (To start the job in five minutes fromnow, set the parameter NEXT_DATE = 5/(24*60) = 1/288.)

VARIABLE jobno NUMBER

EXECUTE DBMS_JOB.SUBMIT(:jobno,'ANALYZE_OBJECT (''TABLE'', ''DEPARTMENTS'');',SYSDATE + 1/288)

PRINT jobno

b. Confirm that the job has been scheduled by using USER_JOBS.

SELECT JOB, NEXT_DATE, NEXT_SEC, WHAT FROM USER_JOBS;



5. Create a procedure called CROSS_AVGSAL that generates a text file report of employees who haveexceeded the average salary of their department. The partial code is provided for you in the filelab14_5.sql.

a. Your program should accept two parameters. The first parameter identifies the output directory. Thesecond parameter identifies the text file name to which your procedure writes.

CREATE OR REPLACE PROCEDURE cross_avgsal

(p_filedir IN VARCHAR2, p_filename1 IN VARCHAR2)

IS

v_fh_1 UTL_FILE.FILE_TYPE;

CURSOR cross_avg IS

SELECT last_name, department_id, salary

FROM employees outer

WHERE salary > (SELECT AVG(salary)

FROM employees inner

GROUP BY outer.department_id)


BEGIN

v_fh_1 := UTL_FILE.FOPEN(p_filedir, p_filename1, 'w');

UTL_FILE.PUTF(v_fh_1,'Employees with more than average salary:\n');

UTL_FILE.PUTF(v_fh_1, 'REPORT GENERATED ON %s\n\n', SYSDATE);

FOR v_emp_info IN cross_avg

LOOP

UTL_FILE.PUTF(v_fh_1, '%s %s \n',

RPAD(v_emp_info.last_name, 30, ' '),

LPAD(TO_CHAR(v_emp_info.salary, '$99,999.00'), 12, ' '));

END LOOP;

UTL_FILE.NEW_LINE(v_fh_1);

UTL_FILE.PUT_LINE(v_fh_1, '*** END OF REPORT ***');

UTL_FILE.FCLOSE(v_fh_1);

END cross_avgsal;

/



b. Your instructor will inform you of the directory location. When you invoke the program,name the second parameter sal_rptxx.txt where xx stands for your user number,such as 01, 15, and so on.

EXECUTE cross_avgsal('$HOME/Utlfile', 'sal_rptxx.txt')(Replace $HOME with the path to the directory Utlfile and xx with your user number)

c. Add an exception handling section to handle errors that may be encountered from usingthe UTL_FILE package.

Sample output from this file follows:

EMPLOYEES OVER THE AVERAGE SALARY OF THEIR DEPARTMENT:REPORT GENERATED ON 26-FEB-01

Hartstein 20 $13,000.00Raphaely 30 $11,000.00Marvis 40 $6,500.00Weiss 50 $8,000.00

...*** END OF REPORT ***

Note: The solution appers on the next page.



CREATE OR REPLACE PROCEDURE cross_avgsal

(p_filedir IN VARCHAR2, p_filename1 IN VARCHAR2)

IS

v_fh_1 UTL_FILE.FILE_TYPE;

CURSOR cross_avg IS

SELECT last_name, department_id, salary

FROM employees outer

WHERE salary > (SELECT AVG(salary)

FROM employees inner

GROUP BY outer.department_id)


BEGIN

v_fh_1 := UTL_FILE.FOPEN(p_filedir, p_filename1, 'w');

UTL_FILE.PUTF(v_fh_1,'Employees with more than average salary:\n');

UTL_FILE.PUTF(v_fh_1, 'REPORT GENERATED ON %s\n\n', SYSDATE);

FOR v_emp_info IN cross_avg

LOOP

UTL_FILE.PUTF(v_fh_1, '%s %s \n',

RPAD(v_emp_info.last_name, 30, ' '),

LPAD(TO_CHAR(v_emp_info.salary, '$99,999.00'), 12, ' '));

END LOOP;

UTL_FILE.NEW_LINE(v_fh_1);

UTL_FILE.PUT_LINE(v_fh_1, '*** END OF REPORT ***');

UTL_FILE.FCLOSE(v_fh_1);

EXCEPTION

WHEN UTL_FILE.INVALID_FILEHANDLE THEN

RAISE_APPLICATION_ERROR (-20001, 'Invalid File.');

UTL_FILE.FCLOSE_ALL;

WHEN UTL_FILE.WRITE_ERROR THEN

RAISE_APPLICATION_ERROR (-20002,

'Unable to write to file');

UTL_FILE.FCLOSE_ALL;

END cross_avgsal;

/


Practice 15 Solutions1. Create a table called PERSONNEL by executing the script file lab15_1.sql. The table contains the

following attributes and data types:

CREATE TABLE personnel

(id NUMBER(6) constraint personnel_id_pk PRIMARY KEY,

last_name VARCHAR2(35),

review CLOB,

picture BLOB);

2. Insert two rows into the PERSONNEL table, one each for employees 2034 and 2035. Use the emptyfunction for the CLOB, and provide NULL as the value for the BLOB.

INSERT INTO personnel

VALUES(2034, 'Allen', EMPTY_CLOB(), NULL);

INSERT INTO personnel

VALUES(2035, 'Bond', EMPTY_CLOB(), NULL);

3. Execute the script lab15_3.sql. The script creates a table named REVIEW_TABLE. This tablecontains annual review information for each employee. The script also contains two statements to insertreview details for two employees.

CREATE TABLE review_table

(employee_id number,

ann_review VARCHAR2(2000));

INSERT INTO review_table

VALUES(2034,'Very good performance this year. Recommended toincrease salary by $500');

INSERT INTO review_table

VALUES(2035,'Excellent performance this year. Recommended toincrease salary by $1000');

COMMIT;

Column Name Data Type Length

ID NUMBER 6

last_name VARCHAR2 35

review CLOB N/A

picture BLOB N/A


Practice 15 Solutions (continued)4. Update the PERSONNEL table.

a. Populate the CLOB for the first row, using the following query in a SQL UPDATE statement:SELECT ann_review

FROM review_table


UPDATE personnel

SET review = (SELECT ann_review

FROM review_table

WHERE employee_id = 2034)

WHERE last_name = 'Allen';

b. Populate the CLOB for the second row, using PL/SQL and the DBMS_LOB package.

Use the following SELECT statement to provide a value:SELECT ann_review

FROM review_table


DECLARE

lobloc CLOB;

text VARCHAR2(2000);

amount NUMBER ;

offset INTEGER;

BEGIN

SELECT ann_review INTO text

FROM review_table

WHERE employee_id =2035;

SELECT review INTO lobloc

FROM personnel

WHERE last_name = 'Bond' FOR UPDATE;

offset := 1;

amount := length(text);

DBMS_LOB.WRITE ( lobloc, amount, offset, text );

END;

/



If you have time...

5. Create a procedure that adds a locator to a binary file into the PICTURE column of the COUNTRIEStable. The binary file is a picture of the country. The image files are named after the country IDs. Youneed to load an image file locator into all rows in Europe region (REGION_ID = 1) in theCOUNTRIES table. The DIRECTORY object name that stores a pointer to the location of the binaryfiles is called COUNTRY_PIC. This object is already created for you.

a. Use the command below to add the image column to the COUNTRIES table. (or runlab15_5_add.sql)

ALTER TABLE countries ADD (picture BFILE);

b. Create a PL/SQL procedure called load_country_image that reads a locator into yourpicture column. Have the program test to see if the file exists, using the functionDBMS_LOB.FILEEXISTS. If the file is not existing, your procedure should display amessage that the file can not be opened. Have your program report information about the loadto the screen.

Note: The solution appears on the next page.

c. Invoke the procedure by passing the name of the directory object COUNTRY_PIC as theparameter. Note that you should pass the directory object in single quotation marks.

EXECUTE load_country_image('COUNTRY_PIC')

Instructor NoteThe class needs to add a BFILE column to the COUNTRIES table to complete question 5 of Practice 15.The script is stored in lab\lab15_5_add.sql.

This practice question 5 requires the instructor to follow the steps in the file lab\lab15_5_setup.sql.The script creates the DIRECTORY objects and grants permissions to these objects. Edit the file and fill inthe database connection for the classroom. Also fill in the file location on the server to identify where theCOUNTRY_PIC and LOG_FILES are located.

The steps in this script must be followed before the students can complete question 5 of Practice 15. Afterrunning the above script, you can verify that the directory alias is created by querying theALL_DIRECTORIES data dictionary view as follows:

SELECT directory_name, directory_path FROM all_directories;



CREATE OR REPLACE PROCEDURE load_country_image

(p_file_loc IN VARCHAR2)

IS

v_file BFILE;

v_filename VARCHAR2(40);

v_record_number NUMBER;

v_file_exists BOOLEAN;

CURSOR country_pic_cursor IS

SELECT country_id

FROM countries

WHERE region_id = 1

FOR UPDATE;

BEGIN

DBMS_OUTPUT.PUT_LINE('LOADING LOCATORS TO IMAGES...');

FOR country_record IN country_pic_cursor

LOOP

v_filename := country_record.country_id || '.tif';

v_file := bfilename(p_file_loc, v_filename);

v_file_exists := (DBMS_LOB.FILEEXISTS(v_file) = 1);

IF v_file_exists THEN

DBMS_LOB.FILEOPEN(v_file);

UPDATE countries

SET picture = bfilename(p_file_loc, v_filename)

WHERE CURRENT OF country_pic_cursor;

DBMS_OUTPUT.PUT_LINE('LOADED LOCATOR TO FILE: '||v_filename|| ' SIZE: ' || DBMS_LOB.GETLENGTH(v_file));

DBMS_LOB.FILECLOSE(v_file);

v_record_number := country_pic_cursor%ROWCOUNT;

ELSE

DBMS_OUTPUT.PUT_LINE('Can not open the file '||v_filename);

END IF;

END LOOP;

DBMS_OUTPUT.PUT_LINE('TOTAL FILES UPDATED: '||v_record_number);

EXCEPTION

WHEN OTHERS THEN

DBMS_LOB.FILECLOSE(v_file);

DBMS_OUTPUT.PUT_LINE('Program Error Occurred: '

|| to_char(SQLCODE) || SQLERRM);

END load_country_image;

/



1. Changes to data are allowed on tables only during normal office hours of 8:45 a.m. until 5:30 p.m.,Monday through Friday.

Create a stored procedure called SECURE_DML that prevents the DML statement from executing outsideof normal office hours, returning the message, “You may make changes only during normal officehours.”

CREATE OR REPLACE PROCEDURE secure_dml

IS

BEGIN

IF TO_CHAR (SYSDATE, 'HH24:MI') NOT BETWEEN '08:45' AND '17:30'

OR TO_CHAR (SYSDATE, 'DY') IN ('SAT', 'SUN')

THEN RAISE_APPLICATION_ERROR (-20205,

'You may make changes only during normal office hours');

END IF;

END secure_dml;

/

2. a. Create a statement trigger on the JOBS table that calls the above procedure.

CREATE OR REPLACE TRIGGER secure_prod

BEFORE INSERT OR UPDATE OR DELETE ON jobs

BEGIN

secure_dml;

END secure_prod;

/

b. Test the procedure by temporarily modifying the hours in the procedure and attempting to insert anew record into the JOBS table. (Example: replace 08:45 with 16:45)

After testing, reset the procedure hours as specified in question 1 and recreate the procedure as inquestion 1 above.

INSERT INTO jobs (job_id, job_title)

VALUES ('HR_MAN', 'Human Resources Manager');



3. Employees should receive an automatic increase in salary if the minimum salary for a job is increased.Implement this requirement through a trigger on the JOBS table.

a. Create a stored procedure named UPD_EMP_SAL to update the salary amount. This procedureaccepts two parameters: the job ID for which salary has to be updated, and the new minimumsalary for this job ID. This procedure is executed from the trigger on the JOBS table.

CREATE OR REPLACE PROCEDURE upd_emp_sal

(p_jobid IN employees.job_id%TYPE,

p_minsal IN employees.salary%TYPE)

IS

BEGIN

UPDATE employees

SET salary = p_minsal

WHERE job_id = p_jobid

AND SALARY < p_minsal;

END upd_emp_sal;

/

b. Create a row trigger named UPDATE_EMP_SALARY on the JOBS table that invokes theprocedure UPD_EMP_SAL when the minimum salary in the JOBS table is updated for aspecified job ID.

CREATE OR REPLACE TRIGGER update_emp_salary

AFTER UPDATE OF min_salary ON jobs

FOR EACH ROWBEGIN

upd_emp_sal(:NEW.job_id, :NEW.min_salary);

END;

/

c. Query the EMPLOYEES table to see the current salary for employees who are programmers.SELECT last_name, first_name, salary

FROM employees

WHERE job_id = 'IT_PROG';


Practice 16 Solutions (continued)d. Increase the minimum salary for the programmer job from 4,000 to 5,000.

UPDATE jobs

SET min_salary = 5000

WHERE job_id = 'IT_PROG';

e. Employee Lorentz (employee ID 107) had a salary of less than 4,500. Verify that her salary hasbeen increased to the new minimum of 5,000.SELECT last_name, first_name, salary

FROM employees


Instructor Note

If you encounter errors while performing practice question 3d, verify that you do not have the triggerstrig_constr_emp and constr_emp_trig in your schema. These triggers are created and droppedwhen you run the demo 16_trig_constr.sql and code example 16_34s.sql. Students did not createthese triggers earlier.


Practice 17 SolutionsA number of business rules that apply to the EMPLOYEES and DEPARTMENTS tables are listed below.

Decide how to implement each of these business rules, by means of declarative constraints or by using triggers.Which constraints or triggers are needed and are there any problems to be expected?

Implement the business rules by defining the triggers or constraints that you decided to create.

A partial package is provided in file lab17_1.sql to which you should add any necessary procedures orfunctions that are to be called from triggers that you may create for the following rules.

(The triggers should execute procedures or functions that you have defined in the package.)

The following code is from the lab17_1.sql file:

REM Package specification with sample procedure specifications

CREATE OR REPLACE PACKAGE mgr_constraints_pkg

IS

PROCEDURE check_president;

PROCEDURE check_mgr;

PROCEDURE new_location(p_deptid IN departments.department_id%TYPE);new_mgr employees.manager_id%TYPE := NULL;

END mgr_constraints_pkg;

/

REM Package Body - fill in the code for the procedures

CREATE OR REPLACE PACKAGE BODY mgr_constraints_pkg

IS

PROCEDURE check_president IS

END check_president;

PROCEDURE check_mgr IS

END check_mgr;

PROCEDURE new_location(p_deptid IN departments.department_id%TYPE)

IS

END new_location;


/


Practice 17 Solutions (continued)The following code is the complete solution for the package specification.

CREATE OR REPLACE PACKAGE mgr_constraints_pkg

IS

PROCEDURE check_president;

PROCEDURE check_mgr;

PROCEDURE new_location

(p_deptid IN departments.department_id%TYPE);

new_mgr employees.manager_id%TYPE := NULL;



Practice 17 Solutions (continued)The following code is the solution for the package body.CREATE OR REPLACE PACKAGE BODY mgr_constraints_pkg

IS

PROCEDURE check_president

IS

v_dummy CHAR(1);

BEGIN

SELECT ‘x’

INTO v_dummy

FROM employees

WHERE job_id = 'AD_PRES';

EXCEPTION


NULL;

WHEN TOO_MANY_ROWS THEN

RAISE_APPLICATION_ERROR(-20001,'President title

already exists');

END check_president;

PROCEDURE check_mgr

IS

count_emps NUMBER := 0;

BEGIN

IF new_mgr IS NOT NULL

THEN

-- count the number of people

-- working for the mgr

SELECT count(*)

INTO count_emps

FROM employees

WHERE manager_id = new_mgr;

END IF;

-- if there are now more than 15,

-- raise an error

IF count_emps > 15

THEN RAISE_APPLICATION_ERROR (-20202,

'Max number of emps exceeded for '|| TO_CHAR(new_mgr));

END IF;

END check_mgr;


Practice 17 Solutions (continued)PROCEDURE new_location

(p_deptid IN departments.department_id%TYPE)

IS


BEGIN

UPDATE employees

SET salary = salary*1.02


END new_location;


/



Business Rules

Rule 1. Sales managers and sales representatives should always receive commission. Employeeswho are not sales managers or sales representatives should never receive a commission.Ensure that this restriction does not validate the existing records of the EMPLOYEEStable. It should be effective only for the subsequent inserts and updates on the table.

Implement rule 1 with a constraint.

ALTER TABLE employees

ADD CONSTRAINT emp_comm_chk

CHECK ((job_id = 'SA_REP' and commission_pct>0) OR

(job_id = 'SA_MAN' and commission_pct>0) OR

(job_id != 'SA_REP' and commission_pct=0))

NOVALIDATE;

Rule 2. The EMPLOYEES table should contain exactly one president.

Test your answer by inserting an employee record with the following details: employee ID400, last name Harris, first name Alice, e-mail ID AHARRIS, job ID AD_PRES, hiredate SYSDATE , salary 20000, and department ID 20.

Note: You do not need to implement a rule for case sensitivity; instead, you need to testfor the number of people with the job title of President.

Implement rule 2 with a trigger.

CREATE OR REPLACE TRIGGER check_pres_title

AFTER INSERT OR UPDATE OF job_id ON employees

BEGIN

mgr_constraints_pkg.check_president;

END check_pres_title;/

INSERT INTO employees

(employee_id, last_name, first_name, email, job_id,

hire_date, salary, department_id)

VALUES (400,'Harris','Alice', 'AHARRIS', 'AD_PRES',

SYSDATE, 20000, 20);



Rule 3. An employee should never be a manager of more than 15 employees.

Test your answer by inserting the following records into the EMPLOYEES table (perform aquery to count the number of employees currently working for manager 100 before insertingthese rows):

i. Employee ID 401, last name Johnson, first name Brian, e-mail ID BJOHNSON, job IDSA_MAN, hire date SYSDATE , salary 11000, manager ID 100, and department ID 80.(This insertion should be successful, because there are only 14 employees working formanager 100 so far.)

ii. Employee ID 402, last name Kellogg, first name Tony, e-mail ID TKELLOG, job IDST_MAN, hire date SYSDATE , salary 7500, manager ID 100, and department ID 50.(This insertion should be unsuccessful, because there are already 15 employees working formanager 100.)


CREATE OR REPLACE TRIGGER set_mgr

AFTER INSERT or UPDATE of manager_id on employees

FOR EACH ROW

BEGIN

-- To get round MUTATING TABLE ERROR

mgr_constraints_pkg.new_mgr := :NEW.manager_id;

END set_mgr;

CREATE OR REPLACE TRIGGER chk_emps

AFTER INSERT or UPDATE of manager_id on employees

BEGIN

mgr_constraints_pkg.check_mgr;

-- if for some reason, SET_MGR is disabled,

-- the global variable is set to null

-- to stop the SELECT COUNT running

mgr_constraints_pkg.new_mgr := NULL;

END chk_emps;

/



hire_date, salary, manager_id, department_id)

VALUES (401,'Johnson','Brian', 'BJOHNSON', 'SA_MAN',

SYSDATE, 11000, 100, 80);


Practice 17 Solutions (continued)SELECT count(*)

FROM employees

WHERE manager_id = 100;



hire_date, salary, manager_id, department_id)

VALUES (402,'Kellogg','Tony', 'TKELLOGG', 'ST_MAN',

SYSDATE, 7500, 100, 50);



Rule 4. Salaries can only be increased, never decreased.

The present salary of employee 105 is 5000. Test your answer by decreasing the salary ofemployee 105 to 4500.


CREATE OR REPLACE TRIGGER check_sal

BEFORE UPDATE OF salary ON employees

FOR EACH ROW

WHEN (NEW.salary < OLD.salary)

BEGIN

RAISE_APPLICATION_ERROR(-20002,'Salary may not be reduced');

END check_sal;

/

UPDATE employees

SET salary = 4500




Rule 5. If a department moves to another location, each employee of that department automaticallyreceives a salary raise of 2 percent.

View the current salaries of employees in department 90.

Test your answer by moving department 90 to location 1600. Query the new salaries of employeesof department 90.


CREATE OR REPLACE TRIGGER change_location

BEFORE UPDATE OF location_id ON departments

FOR EACH ROW

BEGIN

mgr_constraints_pkg.new_location(:OLD.department_id);

END change_location;

/


FROM employees


UPDATE departments

SET location_id = 1600



FROM employees




1. Answer the following questions.

a. Can a table or a synonym be invalid?

A table or a synonym can never be invalidated; however, dependent objects can beinvalidated.

b. Assuming the following scenario, is the stand-alone procedure MY_PROC invalidated?

• The stand-alone procedure MY_PROC depends on the packaged procedureMY_PROC_PACK.

• The MY_PROC_PACK procedure’s definition is changed by recompiling the packagebody.

• The MY_PROC_PACK procedure’s declaration is not altered in the packagespecification.

Although the package body is recompiled, the stand-alone procedure MY_PROC that dependson the packaged procedure MY_PROC_PACK is not invalidated because the packagespecification is not altered

2. Execute the utldtree.sql script. This script is available in your lab folder. Print a tree structureshowing all dependencies involving your NEW_EMP procedure and your VALID_DEPTID function.Query the ideptree view to see your results. (NEW_EMP and VALID_DEPTID were created inlesson 10, “Creating Functions.” You can run the solution scripts for the practice if you need to createthe procedure and function.)

Replace 'your USERNAME' with your username in the following statements.

EXECUTE deptree_fill('PROCEDURE', 'your USERNAME', 'NEW_EMP')

SELECT * FROM ideptree;

EXECUTE deptree_fill('FUNCTION', 'your USERNAME','VALID_DEPTID')

SELECT * FROM ideptree;



If you have time:

3. Dynamically validate invalid objects.

a. Make a copy of your EMPLOYEES table, called EMP_COP.

CREATE TABLE emp_cop AS


b. Alter your EMPLOYEES table and add the column TOTSAL with data type NUMBER(9,2).

ALTER TABLE employees

ADD (totsal NUMBER(9,2));

c. Create a script file to print the name, type, and status of all objects that are invalid.

This is the code that your script file should contain:

SELECT object_name, object_type, status

FROM user_objects

WHERE status = 'INVALID';

d. Create a procedure called COMPILE_OBJ that recompiles all invalid procedures, functions,and packages in your schema.

Make use of the ALTER_COMPILE procedure in the DBMS_DDL package.

CREATE OR REPLACE PROCEDURE compile_obj

IS

CURSOR obj_cur IS

SELECT object_type, object_name

FROM user_objects

WHERE status = 'INVALID'

AND object_type IN ('PROCEDURE', 'FUNCTION', 'PACKAGE',

'PACKAGE BODY')

ORDER BY object_type;

BEGIN

FOR obj_rec IN obj_cur LOOP

DBMS_DDL.ALTER_COMPILE(obj_rec.object_type, user,

obj_rec.object_name);

END LOOP;

END compile_obj;

/

...



Execute the COMPILE_OBJ procedure.

EXECUTE compile_obj

e. Run the script file that you created in question 3c again and check the status columnvalue.

Do you still have INVALID objects? If you do, why are they INVALID?

SELECT object_name, object_type, status

FROM user_objects

WHERE status = 'INVALID';

You may still have invalid objects because the procedure does not take into account objectdependencies.

Table Descriptionsand Data

B

Oracle9i: Program with PL/SQL B-2

ENTITY RELATIONSHIP DIAGRAM


Tables in the Schema

SELECT * FROM tab;


REGIONS Table

DESCRIBE regions

SELECT * FROM regions;


COUNTRIES Table

SELECT * FROM countries;

DESCRIBE countries


LOCATIONS Table

SELECT * FROM locations;

DESCRIBE locations;


DEPARTMENTS Table

DESCRIBE departments

SELECT * FROM departments;


JOBS Table

DESCRIBE jobs

SELECT * FROM jobs;


EMPLOYEES Table

DESCRIBE employees


EMPLOYEES Table

The headings for columns COMMISSION_PCT, MANAGER_ID, and DEPARTMENT_ID are set toCOMM, MGRID, and DEPTID in the following screenshot, to fit the table values across the page.



EMPLOYEES Table (continued)


JOB_HISTORY Table

DESCRIBE job_history

SELECT * FROM job_history;


Creating Program Units by UsingProcedure Builder


65 minutes Total

Oracle9i: Program with PL/SQL C-2

C-2 Copyright © Oracle Corporation, 2001. All rights reserved.

Objectives

After completing this appendix, you should be able todo the following:

• Describe the features of Oracle Procedure Builder

• Manage program units using the Object Navigator

• Create and compile program units using theProgram Unit Editor

• Invoke program units using the PL/SQL Interpreter

• Debug subprograms using the debugger

• Control execution of an interrupted PL/SQLprogram unit

• Test possible solutions at run time

Lesson Aim

You can use different development environments to create PL/SQL program units. In this appendix youlearn to use Oracle Procedure Builder as one of the development environments to create and debugdifferent types of program units. You also learn about the features of the Procedure Builder tool andhow they can be used to create, compile, and invoke subprograms.


PL/SQL Program Constructs

The diagram above displays a variety of different PL/SQL program constructs using the basic PL/SQLblock. In general, a block is either an anonymous block or a named block (subprogram or program unit).

PL/SQL Block Structure

Every PL/SQL construct is composed of one or more blocks. These blocks can be entirely separate ornested within one another. Therefore, one block can represent a small part of another block, which inturn can be part of the whole unit of code.

Note: In the slide, the word “or” prior to the keyword DECLARE is not part of the syntax. It is used inthe diagram to differentiate between starting subprograms and anonymous blocks.

The PL/SQL blocks can be constructed on and use the Oracle server (stored PL/SQL program units).They can also be constructed using the Oracle Developer tools such as Oracle Forms Developer, OracleReport Developer, and so on (application or client-side PL/SQL program units).

Object types are user-defined composite data types that encapsulate a data structure along with thefunctions and procedures needed to manipulate the data. You can create object types either on the Oracleserver or using the Oracle Developer tools.

You can create both application program units and stored program units using Oracle Procedure Builder.Application program units are used in graphical user environment tools such as Oracle Forms. Storedprogram units are stored on the database server and can be shared by multiple applications.


or DECLARE

BEGIN

EXCEPTION

END;

<header>IS|AS

Tools ConstructsAnonymous blocks

Application procedures orfunctions

Application packages

Application triggers

Object types

Database ServerConstructs

Anonymous blocksStored procedures or

functionsStored packages

Database triggers

Object types

PL/SQL Program Constructs



Development Environments

• iSQL*Plus uses the PL/SQL engine in the OracleServer

• Oracle Procedure Builder uses the PL/SQL enginein the client tool or in the Oracle Server. Itincludes:– A GUI development environment for PL/SQL code

– Built-in editors

– The ability to compile, test, and debug code

– Application partitioning that allows drag-and-dropof program units between client and server

iSQL*Plus and Oracle Procedure Builder

PL/SQL is not an Oracle product in its own right. It is a technology employed by the Oracle Serverand by certain Oracle development tools. Blocks of PL/SQL are passed to, and processed by, a PL/SQLengine. That engine may reside within the tool or within the Oracle Server.

There are two main development environments for PL/SQL: iSQL*Plus and Oracle Procedure Builder.This course covers creating program units using iSQL*Plus.

About Procedure Builder

Oracle Procedure Builder is a tool you can use to create, execute, and debug PL/SQL programs used inyour application tools, such as a form or report, or on the Oracle server through its graphical interface.

Integrated PL/SQL Development Environment

Procedure Builder's development environment contains a build-in editor for you to create or editsubprograms. You can compile, test, and debug your code.

Unified Client-Server PL/SQL Development

Application partitioning through Procedure Builder is available to assist you with distribution of logicbetween client and server. Users can drag and drop a PL/SQL program unit between the client and theserver.



Developing Procedures and FunctionsUsing iSQL*Plus

Using iSQL*Plus

Use a text editor to create a script to define your procedure or function. Browse and upload the scriptinto the iSQL*Plus input window. Execute the script by clicking the EXECUTE button.

The example in the slide creates a stored procedure without any parameters. The procedure recordsthe username and current date in a database table.



Developing Procedures and Functions UsingOracle Procedure Builder

Start Procedure Builder from Windows

Procedure Builder contains object navigator where you can see all the program units that you created.You can open, edit, compile, debug, and save the program units by using a graphical editor.

Instructor NoteDemonstration: Cut and paste code from 09_logexec.sql.

Purpose: To show how to create a procedure using Procedure Builder.

1. Invoke Procedure Builder.

2. Click Program Units in the Object Navigator pane.

3. Click the Create button.

4. Cut and paste code from 09_logexec.sql into the Program Unit editor. Do not includeCREATE OR REPLACE, or the forward slash, in the code.

5. Click the Compile button.



Components ofProcedure Builder

Component

Object Navigator

PL/SQL Interpreter

Program Unit Editor

Stored ProgramUnit Editor

Database Trigger Editor

Function

Manages PL/SQL constructs;performs debug actions

Debugs PL/SQL code; evaluatesPL/SQL code in real time

Creates and edits PL/SQL sourcecode

Creates and edits server-sidePL/SQL source code

Creates and edits database triggers

Components of Procedure Builder

Procedure Builder is an integrated development environment. It enables you to edit, compile, test, anddebug client-side and server-side PL/SQL program units within a single tool.

The Object Navigator

The Object Navigator provides an outline-style interface to browse objects, view the relationshipsbetween them, and edit their properties.

The Interpreter Pane

The Interpreter pane is the central debugging workspace of the Oracle Procedure Builder. It is a windowwith two regions where you display, debug, and run PL/SQL program units. It also interactivelysupports the evaluation of PL/SQL constructs, SQL commands, and Procedure Builder commands.

The Program Unit Editor

The easiest and most common place to enter PL/SQL source code is in the Program Unit Editor. Youcan use it to edit, compile, and browse warning and error messages during application development. TheStored Program Unit Editor is a GUI environment for editing server-side packages and subprograms.The compile operation submits the source text to the server-side PL/SQL compiler.

The Database Trigger Editor

The Database Trigger Editor is a GUI environment for editing database triggers. The compile operationsubmits the source text to the server-side PL/SQL compiler.



Developing Program Unitsand Stored Programs Units

Stored program unitsin the Oracle server

ProcedureBuilder

Program unitsin a PL/SQL library

Server-sidecode

Client-sidecode

Program Units and Stored Program Units

Use Procedure Builder to develop PL/SQL subprograms that can be used by client and serverapplications.

Program units are client-side PL/SQL subprograms that you use with client applications, such as OracleDeveloper. Stored program units are server-side PL/SQL subprograms that you use with allapplications, client or server.

Developing PL/SQL Code

Client-side code:

• Create program units by using the Program Unit Editor

• Drag a server-side subprogram to the client by using the Object Navigator

Server-side code:

• Create stored programs by using the Stored Program Unit Editor

• Drag a client-side program unit to the server by using the Object Navigator


Components of the Object Navigator

The following descriptions correspond to the numbered components on the slide:

1. Location indicator: Shows your current location in the hierarchy.

2. Subobject indicator: Allows you to expand and collapse nodes to view or hide object information.Different icons represent different classes of objects.

3. Type icon: Indicates the type of object, followed by the name of the object. In the example, theicon indicates that LOG_EXECUTION is a PL/SQL block. If you double-click the icon, ProcedureBuilder opens the Program Unit Editor and displays the code of that object.

4. Object name: Shows you the names of the objects.

5. Find field: Allows you to search for objects.

Object Navigator

The Object Navigator is Procedure Builder's browser for locating and working with both client andserver program units, libraries, and triggers.

The Object Navigator allows you to expand and collapse nodes, cut and paste, search for an object, anddrag PL/SQL program units between the client and the server side.


Procedure Builder Components:The Object Navigator

12

3 4

5


Components of the Object Navigator: Vertical Button Bar

The vertical button bar on the Object Navigator provides convenient access for many of the actionsfrequently performed from the File, Edit, and Navigator menus.

1. Open: Opens a library from the file system or from the Oracle server.

Save: Saves a library in the file system or on the Oracle server.

2. Cut: Cuts the selected object and stores it in the clipboard. Cutting an object also cuts any objectsowned by that object.

Copy: Makes a copy of the selected object and stored it in the clipboard. Copying an object alsocopies any objects owned by that object.

Paste: Pastes the cut or copied module into the selected location. Note that objects must be copiedto a valid location in the object hierarchy.

3. Create: Creates a new instance of the currently selected object.

Delete: Deletes the selected object with confirmation.

4. Expand, Collapse, Expand All, and Collapse All: Expands or collapses one or all levels ofsubobjects of the currently selected object.


Procedure Builder Components:The Object Navigator

1

2

3

4


Objects of the Object Navigator

By using the Object Navigator, you can display a hierarchical listing of all objects you have access toduring your session.


Procedure Builder Components:Objects of the Navigator

• Program Units– Specification

– References

– Referenced By

• Libraries

• Attached Libraries

• Built-in Packages

• Debug Actions

• Stack

• Database Objects

Object Nodes DescriptionProgram Units PL/SQL constructs that can be independently recognized and processed by the

PL/SQL compiler.Program Units:Specification

Name, parameter, and return type (functions only) of the program unit.

Program Units:References

Procedures, functions, anonymous blocks, and tables that the program unitreferences.

Program Units:Referenced By

Procedures, functions, anonymous blocks, and tables that reference theprogram unit.

Libraries Collection of PL/SQL packages, procedures, and functions stored in thedatabase or the file system.

AttachedLibraries

Referenced libraries stored in the database or the file system.

Built-inPackages

PL/SQL constructs that can be referenced while debugging program units.

Debug Actions Actions that enable you to monitor or interrupt the execution of PL/SQLprogram units.

Stack Chain of subprogram calls, from the initial entry point down to the currentlyexecuting subprogram.

DatabaseObjects

Collection of server-side stored program units, libraries, tables, and views.



Developing Stored Procedures

OracleProcedure

Builder

Code

Compile and Save

Source code

Execute

P code

Oracle

How to Develop Stored Program Units

Use the following steps to develop a stored program unit:

1. Enter the syntax in the Program Unit editor.

2. Click the Save button to compile and save the code.

The source code is compiled into P code.

Instructor Note

The PL/SQL compiler generates the pseudocode or P code, based on the parsed code. The PL/SQLengine executes this code when the procedure is invoked.



Procedure Builder Components:The Program Unit Editor

12

3

Program Unit Editor

The following descriptions correspond to the numbered components on the slide:

1. Compile, Apply, Revert, New, Delete, Close, and Help buttons

2. Name drop-down list

3. Source text pane

Program Unit Editor

Use the Program Unit Editor to edit, compile, and browse warning and error messages duringdevelopment of client-side PL/SQL subprograms.

To bring a subprogram into the source text pane, select an option from the Name drop-down list. Use thebuttons to decide which action to take once you are in the Program Unit Editor.



Procedure Builder Components:The Stored Program Unit Editor

The Stored Program Unit Editor

Use the Stored Program Unit Editor to edit server-side PL/SQL constructs. The Save operation submitsthe source text to the server-side PL/SQL compiler.



1

2

Creating a Client-SideProgram Unit

3

4

5

How to Create a Client-Side Program Unit

1. Select the Program Units object or subobject.

2. Click the Create button. The New Program Unit dialog box appears.

3. Enter the name of your subprogram, select the subprogram type, and click the OK button to acceptthe entries.

4. The Program Unit editor is displayed. It contains the skeleton for your PL/SQL construct. Thecursor is automatically positioned on the line beneath the BEGIN keyword. You can now write thecode.

5. When you finish writing the code, click Compile in the Program Unit Editor.

Error messages generated during compilation are displayed in the compilation message pane in theProgram Unit window. When you select an error message, the cursor moves to the location of theerror in the program screen.

When your PL/SQL code is error free, the compilation message disappears, and the SuccessfullyCompiled message appears in the status line of the Program Unit Editor.

Note: Program units that reside in the Program Units node are lost when you exit Procedure Builder.You must export them to a file, save them in a PL/SQL library, or store them in the database.


How to Create a Server-Side Program Unit

1. Select the Database Objects node in the Object Navigator, expand the schema name, and click StoredProgram Units.

2. Click Create.

3. In the New Program Unit window, enter the name of the subprogram, select the subprogram type,and click OK to accept the entries.

4. The Stored Program Unit editor is displayed. It contains the skeleton for your PL/SQL construct. Thecursor is automatically positioned on the line beneath the BEGIN keyword. You can now write thecode.

5. When you finish writing the code, click Save in the Stored Program Unit Editor.

Error messages generated during compilation are displayed in a compilation message at the bottomof the window. Click an error message to move to the location of the error.

When the PL/SQL code is error-free, the compilation message does not appear. The SuccessfullyCompiled message appears in the status line at the bottom of the Stored Program Unit Editorwindow.


Creating a Server-SideProgram Unit

4

5

1

2

3


Application Partitioning

Using Procedure Builder you can create PL/SQL program units on both the client and the server. Youcan also use Procedure Builder to copy program units created on the client into stored program units onthe server (or vice versa). You can do this by a dragging the program unit to the destination StoredProgram Units node in the appropriate schema.

PL/SQL code that is stored in the server is processed by the server-side PL/SQL engine; therefore, anySQL statements contained within the program unit do not have to be transferred between a clientapplication and the server.

Program units on the server are potentially accessible to all applications (subject to user securityprivileges).


Transferring Program Units BetweenClient and Server



Procedure Builder Components:The PL/SQL Interpreter

1

2

3

Components of the PL/SQL Interpreter

1. Source pane: Displays the PL/SQL code of your program.

2. Navigator pane: Displays the same information as the Object Navigator, but within the PL/SQLInterpreter.

3. Interpreter pane: Allows you to execute subprograms, Procedure Builder commands, and SQLstatements.

To execute subprograms, enter the name of your PL/SQL program at the PL/SQL prompt, provide anyparameters, and terminate with a semicolon.

PL/SQL> construct_name [parameter1|parameter2,…];

To execute SQL statements, enter your SQL statement and terminate with a semicolon.

PL/SQL> SELECT *

+> FROM departments;

Instructor Note

To demonstrate this concept, use the View menu option, and then select which of the panes you wish tosee at any time. There must always be one pane on display.



Creating Client-Side Program Units

How to Create a Client-Side Program Units

1. Select the Program Units node in the Object Navigator.

2. Click Create. The New Program Unit dialog box appears.

3. Enter a name for the procedure. Note that the default program unit type is Procedure. Click OK toaccept these entries. The program unit name appears in the Object Navigator.

– The Program Unit editor appears, containing the procedure name and IS, BEGIN, and ENDstatements.

– The cursor is automatically positioned on the line beneath the BEGIN keyword.

4. Enter the source code.

5. Click Compile. Error messages generated during compilation are displayed in the compilationmessage pane (the lower half of the window).

6. Select an error message to go to the location of the error in the source text pane.

When successfully compiled, a message is displayed in the lower right hand corner of the ProgramUnit Editor window.

7. Save the source code in a file (M) File > Export.

Note: The keywords CREATE, and CREATE OR REPLACE and the forward slash are invalid inProcedure Builder.



Creating Server-Side Program Units

CreateDelete

How to Create a Server-Side Program Units

1. Select File > Connect. Then enter your username, password, and database connect string.

2. Expand the Database Objects node in the Object Navigator.

3. Expand your schema name.

4. Click the Stored Program Units node under that schema.

5. Click Create in the Object Navigator.

6. Enter the name for the procedure in the New Program Unit dialog box.

7. Click OK to accept.

8. Enter the source code and click save.

Note: The keywords CREATE, and CREATE OR REPLACE and the forward slash are invalid inProcedure Builder.



The DESCRIBE Command inProcedure Builder

Describing Procedures and FunctionsTo display a procedure or function, its parameter list, and other information, use the .DESCRIBEcommand in Procedure Builder.

Example

Display information about the FORMAT_PHONE procedure.



Listing Code of Stored Program Units

ExpandandCollapsebuttons

Storedprocedureicon

Listing Code of a Stored Procedure

1. Select File > Connect and enter your username, password, and database.

2. Select Database Objects and click the Expand button.

3. Select the schema of the procedure owner and click the Expand button.

4. Select Stored Program Units and click the Expand button.

5. Double-click the icon of the stored procedure. The Stored Program Unit editor appears in thewindow and contains the code of the procedure.

The ADD_DEPT Procedure Code

The example above shows the PL/SQL Program Unit editor with the code for the ADD_DEPTprocedure.

The code can now be saved to a file.

1. Select File > Export and enter the name of your file in the Open dialog box.

2. Click OK. A file containing your stored procedure text (.pls extension) is created.



Navigating Compilation Errorsin Procedure Builder

How to Resolve Compilation Errors

1. Click Compile.

2. Select an error message.

The cursor moves to the location of the error in the source pane.

3. Resolve the syntax error and click Compile.



Procedure Builder Built-in Package:TEXT_IO

• The TEXT_IO package:

– Contains a procedure PUT_LINE, which writesinformation to the PL/SQL Interpreter window

– Is used for client-side program units

• The TEXT_IO.PUT_LINE accepts one parameter

PL/SQL> TEXT_IO.PUT_LINE(1);1

TEXT_IO Built-in Package

You can use TEXT_IO packaged procedures to output values and messages from a client-sideprocedure or function to the PL/SQL Interpreter window.

TEXT_IO is a built-in package that is part of Procedure Builder.

Use the Oracle supplied package DBMS_OUTPUT to debug server-side procedures, and the ProcedureBuilder built-in, TEXT_IO, to debug client-side procedures.

Note:

• You cannot use TEXT_IO to debug server-side procedures. The program will fail to compilesuccessfully because TEXT_IO is not stored in the database.

• DBMS_OUTPUT does not display messages in the PL/SQL Interpreter window if you execute aprocedure from Procedure Builder.



Executing Functions inProcedure Builder: Example

Display the tax based on a specified value.

PL/SQL> .CREATE NUMBER x PRECISION 4PL/SQL> :x := tax(1000);PL/SQL> TEXT_IO.PUT_LINE (TO_CHAR(:x));80

Calling environment TAX function

v_value1000

RETURN (computed value)

ExampleExecute the TAX function from Procedure Builder:

1. Create a host variable to hold the value returned from the function. Use the .CREATE syntax atthe Interpreter prompt.

2. Create a PL/SQL expression to invoke the function TAX, passing a numeric value to the function.Note the use of the colon (:) to reference a host variable.

3. View the result of the function call by using the PUT_LINE procedure in the TEXT_IO package.



Creating Statement Triggers

How to Create a Statement Trigger When Using Procedure BuilderYou can also create the same BEFORE statement trigger in Procedure Builder.

1. Connect to the database.

2. Click the Database Objects node in the Object Navigator.

3. Select the Database Trigger editor from the Program menu.

4. Select a table owner and a table from the Table owner and Table drop-down lists.

5. Click New to start creating the trigger.

6. Select one of the Triggering option buttons to choose the timing component.

7. Select Statement to choose the event component.

8. In the Trigger Body region, enter the trigger code.

9. Click Save. Your trigger code will now be compiled by the PL/SQL engine in the server. Oncesuccessfully compiled, your trigger is stored in the database and automatically enabled.

Note: If the trigger has compilation errors, the error message appears in a separate window.



Creating Row Triggers

How to Create a Row Trigger When Using Procedure BuilderYou can also create the same BEFORE row trigger in Procedure Builder.


2. Click the Database Objects node in the Object Navigator.

3. Select the Database Trigger Editor from the Program menu.

4. Select a table owner and a table from the corresponding drop-down lists.

5. Click New to start creating the trigger.

6. Select the Triggering option button to choose the timing component.

7. Select the appropriate Statement check boxes to choose the events component.

8. In the For Each region, select the Row option button to designate the trigger as a row trigger.

9. Complete the Referencing OLD As and NEW As fields if you want to modify the correlationnames. In the When field, enter a WHEN condition to restrict the execution of the trigger. Thesefields are optional and are available only with row triggers.

10. Enter the trigger code.

11. Click Save. The trigger code is now compiled by the PL/SQL engine in the server. Whensuccessfully compiled, the trigger is stored in the database and automatically enabled.



Removing Server-Side Program Units

Using Procedure Builder:


2. Expand the Database Objects node.

3. Expand the schema of the owner of the program unit.

4. Expand the Stored Program Units node.

5. Click the program unit that you want to drop.

6. Click Delete in the Object Navigator.

7. Click Yes to confirm.

Removing a Server-Side Program Unit

When you decide to delete a stored program unit, an alert box displays with the following message:

"Do you really want to drop stored program unit <program unit name>?". Click Yes to drop the unit.

In the Stored Program Units Editor, you can also click DROP to remove the procedure from the server.



Removing Client-SideProgram Units

Using Procedure Builder:

1. Expand the Program Units node.

2. Click the program unit that you want to remove.

3. Click Delete in the Object Navigator.

4. Click Yes to confirm.

Removing a Client-Side Program Unit

Follow the steps in the slide to remove a procedure from Procedure Builder.

If you have exported the code that built your procedure to a text file and you want to delete that file fromthe client, you must use the appropriate operating system command.



Debugging Subprograms by UsingProcedure Builder

Debugging Subprograms by Using Procedure Builder

You can perform debug actions on a server-side or client-side subprogram using Procedure Builder.Use the following steps to load the subprogram:

1. From the Object Navigator, select Program > PL/SQL Interpreter.

2. In the menu, select View > Navigator Pane.

3. From the Navigator pane, expand either the Program Units or the Database objects node.

4. Locate the program unit that you want to debug and click it.



Listing Code in the Source Pane

1

2

3

Listing Code in the Source Pane

Performing Debug Actions in the Interpreter

You can use the Object Navigator to examine and modify parameters in an interrupted program. Byinvoking the Object Navigator within the Interpreter, you can perform debugging actions entirely withinthe Interpreter window. Alternatively, you can interact with the Object Navigator and Interpreterwindows separately.

1. Invoking the Object Navigator Pane

– Select PL/SQL Interpreter from the Tools menu to open the Interpreter if it is not alreadyopen.

– Select Navigator Pane from the View menu.

– The Navigator pane is inserted between the Source and the Interpreter panes.

– Drag the split bars to adjust the size of each pane.

2. Listing Source Text in the Source Pane

– Click the Program Units node in the Navigator pane to expand the list.

The list of program units is displayed.

– Click the object icon of the program unit to be listed.

3. The source code is listed in the Source pane of the Interpreter.



1

2

Setting a Breakpoint


If you encounter errors while compiling or running your application, you should test the code anddetermine the cause for the error. To determine the cause of the error effectively, review the code, lineby line. Eventually, you should identify the exact line of code causing the error. You can use abreakpoint to halt execution at any given point and to permit you to examine the status of the code on aline-by-line basis.


1. Double click the executable line of code on which to break. A "B(n)" is placed in the line wherethe break is set.

2. The message Breakpoint #n installed at line i of name is shown in theInterpreter pane.

Note: Breakpoints also can be set using debugger commands in the Interpreter pane. Test breakpoints byentering the program unit name at the Interpreter PL/SQL prompt.

Monitoring Debug Actions

Debug actions, like breakpoints, can be viewed in the Object Navigator under the heading DebugActions. Double-click the Debug Actions icon to view a description of the breakpoint. Removebreakpoints by double-clicking the breakpoint line number



Debug Commands

Step Over

StepInto

Step Out

Reset

Go

Debug Commands

Reviewing Code

When a breakpoint is reached, you can use a set of commands to step through the code. You canexecute these commands by clicking the command buttons on the Interpreter toolbar or by enteringthe command at the Interpreter prompt.

Commands for Stepping through Code

Command Description

Step Into Advances execution into the next executable line of code

Step Over Bypasses the next executable line of code and advances to thesubsequent line

Step Out Resumes to the end of the current level of code, such as thesubprogram

Go Resumes execution until either the program unit ends or isinterrupted again by a debug action

Reset Aborts the execution at the current levels of debugging



Stepping through Code

1

2

3

Stepping Through Code

Determining the Cause of Error

After the breakpoint is found at run time, you can begin stepping through the code. An arrow (=>)indicates the next line of code to execute.

1. Click the Step Into button.

2. A single line of code is executed. The arrow moves to the next line of code.

3. Repeat step 1 as necessary until the line causing the error is found.

The arrow continues to move forward until the erroneous line of code is found. At that time, PL/SQLdisplays an error message.



Changing a Value

1

2

3

4

Changing a Value

Examining Local Variables

Using Procedure Builder, you can examine and modify local variables and parameters in an interruptedprogram. Use the Stack node in the Navigator pane to view and change the values of local variables andparameters associated with the current program unit located in the call stack. When debugging code,check for the absence of values as well as incorrect values.

Examining Values and Testing the Possible Solution

1. Click the Stack node in the Object Navigator or Navigator pane to expand it.

2. Clock the value of the variable to edit. For example, select variable 1.

The value 1 becomes an editable field.

3. Enter the new value and click anywhere in the Navigator pane to end the variable editing, forexample, enter 3.

The following statement is displayed in the Interpreter pane:

(debug1) PL/SQL> debug.seti('I',3);

4 Click the Go button to resume execution through the end of the program unit.

Note: Variables and parameters can also be changed by using commands at the Interpreter PL/SQLprompt.



Summary

In this appendix, you should have learned how to:

• Use Procedure Builder:– Application partitioning

– Built-in editors

– GUI execution environment

• Describe the components of Procedure Builder– Object Navigator

– Program Unit Editor

– PL/SQL Interpreter

– Debugger


REF Cursors

Oracle9i: Program with PL/SQL D-2

Copyright © Oracle Corporation, 2001. All rights reserved.D-2

Cursor Variables

• Cursor variables are like C or Pascal pointers,which hold the memory location (address) of anitem instead of the item itself

• In PL/SQL, a pointer is declared as REF X, whereREF is short for REFERENCE and X stands for aclass of objects

• A cursor variable has the data type REF CURSOR

• A cursor is static, but a cursor variable is dynamic

• Cursor variables give you more flexibility

Cursor Variables

Cursor variables are like C or Pascal pointers, which hold the memory location (address) of some iteminstead of the item itself. Thus, declaring a cursor variable creates a pointer, not an item. In PL/SQL, apointer has the datatype REF X, where REF is short for REFERENCE and X stands for a class ofobjects. A cursor variable has datatype REF CURSOR.

Like a cursor, a cursor variable points to the current row in the result set of a multirow query.However, cursors differ from cursor variables the way constants differ from variables. A cursor isstatic, but a cursor variable is dynamic because it is not tied to a specific query. You can open a cursorvariable for any type-compatible query. This gives you more flexibility.

Cursor variables are available to every PL/SQL client. For example, you can declare a cursor variablein a PL/SQL host environment such as an OCI or Pro*C program, and then pass it as an input hostvariable (bind variable) to PL/SQL. Moreover, application development tools such as Oracle Formsand Oracle Reports, which have a PL/SQL engine, can use cursor variables entirely on the client side.The Oracle server also has a PL/SQL engine. You can pass cursor variables back and forth between anapplication and server through remote procedure calls (RPCs).



Why Use Cursor Variables?

• You can use cursor variables to pass query resultsets between PL/SQL stored subprograms andvarious clients.

• PL/SQL can share a pointer to the query work areain which the result set is stored.

• You can pass the value of a cursor variable freelyfrom one scope to another.

• You can reduce network traffic by having aPL/SQL block open (or close) several host cursorvariables in a single round trip.

Why Use Cursor Variables?

You use cursor variables to pass query result sets between PL/SQL stored subprograms and variousclients. Neither PL/SQL nor any of its clients owns a result set; they simply share a pointer to thequery work area in which the result set is stored. For example, an OCI client, an Oracle Formsapplication, and the Oracle server can all refer to the same work area.

A query work area remains accessible as long as any cursor variable points to it. Therefore, you canpass the value of a cursor variable freely from one scope to another. For example, if you pass a hostcursor variable to a PL/SQL block that is embedded in a Pro*C program, the work area to which thecursor variable points remains accessible after the block completes.

If you have a PL/SQL engine on the client side, calls from client to server impose no restrictions. Forexample, you can declare a cursor variable on the client side, open and fetch from it on the server side,then continue to fetch from it back on the client side. Also, you can reduce network traffic by having aPL/SQL block open (or close) several host cursor variables in a single round trip.

A cursor variable holds a reference to the cursor work area in the PGA instead of addressing it with astatic name. Because you address this area by a reference, you gain the flexibility of a variable.



Defining REF CURSOR Types

• Define a REF CURSOR type.

Define a REF CURSOR typeTYPE ref_type_name IS REF CURSOR [RETURN return_type];

• Declare a cursor variable of that type.

ref_cv ref_type_name;

• Example:DECLARETYPE DeptCurTyp IS REF CURSOR RETURNdepartments%ROWTYPE;dept_cv DeptCurTyp;

Defining REF CURSOR TypesTo define a REF CURSOR, you perform two steps. First, you define a REF CURSOR type, and thenyou declare cursor variables of that type. You can define REF CURSOR types in any PL/SQL block,subprogram, or package using the following syntax:

TYPE ref_type_name IS REF CURSOR [RETURN return_type];

in which:

ref_type_name is a type specifier used in subsequent declarations of cursor variables

return_type represents a record or a row in a database table

In the following example, you specify a return type that represents a row in the database tableDEPARTMENT.

REF CURSOR types can be strong (restrictive) or weak (nonrestrictive). As the next example shows, astrong REF CURSOR type definition specifies a return type, but a weak definition does not:

DECLARE

TYPE EmpCurTyp IS REF CURSOR RETURN employees%ROWTYPE; -- strong

TYPE GenericCurTyp IS REF CURSOR; -- weak


Defining REF CURSOR Types (continued)

Strong REF CURSOR types are less error prone because the PL/SQL compiler lets you associate astrongly typed cursor variable only with type-compatible queries. However, weak REF CURSORtypes are more flexible because the compiler lets you associate a weakly typed cursor variable withany query.

Declaring Cursor VariablesAfter you define a REF CURSOR type, you can declare cursor variables of that type in any PL/SQLblock or subprogram. In the following example, you declare the cursor variable DEPT_CV:

DECLARE

TYPE DeptCurTyp IS REF CURSOR RETURN departments%ROWTYPE;

dept_cv DeptCurTyp; -- declare cursor variable

Note: You cannot declare cursor variables in a package. Unlike packaged variables, cursor variablesdo not have persistent states. Remember, declaring a cursor variable creates a pointer, not an item.Cursor variables cannot be saved in the database; they follow the usual scoping and instantiation rules.

In the RETURN clause of a REF CURSOR type definition, you can use %ROWTYPE to specify arecord type that represents a row returned by a strongly (not weakly) typed cursor variable, as follows:

DECLARE

TYPE TmpCurTyp IS REF CURSOR RETURN employees%ROWTYPE;

tmp_cv TmpCurTyp; -- declare cursor variable

TYPE EmpCurTyp IS REF CURSOR RETURN tmp_cv%ROWTYPE;

emp_cv EmpCurTyp; -- declare cursor variable

Likewise, you can use %TYPE to provide the datatype of a record variable, as the following exampleshows:

DECLARE

dept_rec departments%ROWTYPE; -- declare record variable

TYPE DeptCurTyp IS REF CURSOR RETURN dept_rec%TYPE;

dept_cv DeptCurTyp; -- declare cursor variable

In the final example, you specify a user-defined RECORD type in the RETURN clause:

DECLARE

TYPE EmpRecTyp IS RECORD (

empno NUMBER(4),

ename VARCHAR2(1O),

sal NUMBER(7,2));

TYPE EmpCurTyp IS REF CURSOR RETURN EmpRecTyp;

emp_cv EmpCurTyp; -- declare cursor variable

Cursor Variables As Parameters

You can declare cursor variables as the formal parameters of functions and procedures. In thefollowing example, you define the REF CURSOR type EmpCurTyp, and then declare a cursorvariable of that type as the formal parameter of a procedure:

DECLARE

TYPE EmpCurTyp IS REF CURSOR RETURN emp%ROWTYPE;

PROCEDURE open_emp_cv (emp_cv IN OUT EmpCurTyp) IS ...



Using the OPEN-FOR, FETCH, and CLOSEStatements

• The OPEN-FOR statement associates a cursorvariable with a multirow query, executes thequery, identifies the result set, and positions thecursor to point to the first row of the result set.

• The FETCH statement returns a row from the resultset of a multirow query, assigns the values ofselect-list items to corresponding variables orfields in the INTO clause, increments the countkept by %ROWCOUNT, and advances the cursor tothe next row.

• The CLOSE statement disables a cursor variable.

Using the OPEN-FOR, FETCH, and CLOSE Statements

You use three statements to process a dynamic multirow query: OPEN-FOR, FETCH, and CLOSE.First, you OPEN a cursor variable FOR a multirow query. Then, you FETCH rows from the result setone at a time. When all the rows are processed, you CLOSE the cursor variable.

Opening the Cursor Variable

The OPEN-FOR statement associates a cursor variable with a multirow query, executes the query,identifies the result set, positions the cursor to point to the first row of the results set, then sets therows-processed count kept by %ROWCOUNT to zero. Unlike the static form of OPEN-FOR, thedynamic form has an optional USING clause. At run time, bind arguments in the USING clausereplace corresponding placeholders in the dynamic SELECT statement. The syntax is:

OPEN {cursor_variable | :host_cursor_variable} FOR dynamic_string

[USING bind_argument[, bind_argument]...];

where CURSOR_VARIABLE is a weakly typed cursor variable (one without a return type),HOST_CURSOR_VARIABLE is a cursor variable declared in a PL/SQL host environment such as anOCI program, and dynamic_string is a string expression that represents a multirow query.


Using the OPEN-FOR, FETCH, and CLOSE Statements (continued)In the following example, the syntax declares a cursor variable, and then associates it with a dynamicSELECT statement that returns rows from the EMPLOYEES table:DECLARETYPE EmpCurTyp IS REF CURSOR; -- define weak REF CURSOR typeemp_cv EmpCurTyp; -- declare cursor variablemy_ename VARCHAR2(15);my_sal NUMBER := 1000;BEGINOPEN emp_cv FOR -- open cursor variable

'SELECT last_name, salary FROM employees WHERE salary > :s'USING my_sal;

...END;

Any bind arguments in the query are evaluated only when the cursor variable is opened. Thus, to fetchrows from the cursor using different bind values, you must reopen the cursor variable with the bindarguments set to their new values.

Fetching from the Cursor VariableThe FETCH statement returns a row from the result set of a multirow query, assigns the values ofselect-list items to corresponding variables or fields in the INTO clause, increments the count kept by%ROWCOUNT, and advances the cursor to the next row. Use the following syntax:FETCH {cursor_variable | :host_cursor_variable}

INTO {define_variable[, define_variable]... | record};

Continuing the example, fetch rows from cursor variable EMP_CV into define variables MY_ENAMEand MY_SAL:LOOP

FETCH emp_cv INTO my_ename, my_sal; -- fetch next row

EXIT WHEN emp_cv%NOTFOUND; -- exit loop when last row is fetched

-- process row

END LOOP;

For each column value returned by the query associated with the cursor variable, there must be acorresponding, type-compatible variable or field in the INTO clause. You can use a different INTOclause on separate fetches with the same cursor variable. Each fetch retrieves another row from thesame result set. If you try to fetch from a closed or never-opened cursor variable, PL/SQL raises thepredefined exception INVALID_CURSOR.

Closing the Cursor VariableThe CLOSE statement disables a cursor variable. After that, the associated result set is undefined. Usethe following syntax:CLOSE {cursor_variable | :host_cursor_variable};

In this example, when the last row is processed, close cursor variable EMP_CV:LOOP

FETCH emp_cv INTO my_ename, my_sal;

EXIT WHEN emp_cv%NOTFOUND;

-- process row

END LOOP;

CLOSE emp_cv; -- close cursor variable

If you try to close an already-closed or never-opened cursor variable, PL/SQL raisesINVALID_CURSOR.



An Example of Fetching

DECLARETYPE EmpCurTyp IS REF CURSOR;emp_cv EmpCurTyp;emp_rec employees%ROWTYPE;sql_stmt VARCHAR2(200);my_job VARCHAR2(10) := 'ST_CLERK';

BEGINsql_stmt := 'SELECT * FROM employees

WHERE job_id = :j';OPEN emp_cv FOR sql_stmt USING my_job;LOOP

FETCH emp_cv INTO emp_rec;EXIT WHEN emp_cv%NOTFOUND;-- process record

END LOOP;CLOSE emp_cv;

END;/

An Example of Fetching

The example in the preceding slide shows that you can fetch rows from the result set of a dynamicmultirow query into a record. First you must define a REF CURSOR type, EmpCurTyp. Next youdefine a cursor variable emp_cv, of the type EmpcurTyp. In the executable section of the PL/SQLblock, the OPEN-FOR statement associates the cursor variable EMP_CV with the multirow query,sql_stmt. The FETCH statement returns a row from the result set of a multirow query and assignsthe values of select-list items to EMP_REC in the INTO clause. When the last row is processed, closethe cursor variable EMP_CV.

Oracle9i: Program with PL/SQL Index-3

A

actual parameter 2-6

anonymous blocks 1-7

application trigger 9-3

aUTHID CURRENT_USE 4-5

B

BEFORE statement trigger 9-14

BEGIN 1-7

benefits 2-26

BFILE 8-12

BFILENAME 8-12

binding 7-5

C

CALL statement 10-7

CLOSE_CONNECTION 7-31

CREATE PROCEDURE 2-5

CREATE ANY DIRECTORY 8-13

D

database trigger 9-3, 10-11

DBA_JOB 7-19

DBA_JOBS_RUNNING 7-19

DBMS_DDL 7-12

DBMS_JOB 7-13

DBMS_JOB.BROKEN 7-18

DBMS_JOB.REMOVE 7-18

DBMS_JOB.RUN 7-18

DBMS_LOB 7-21, 8-12

data dictionary view 4-9, 4-11

data type 3-4

DBMS_OUTPUT 4-16

DBMS_SQL 7-6

DECLARE 1-7

definer's-rights 4-4

DEPTREE 11-8

DIRECTORY 8-10

DROP PROCEDURE 2-25

dynamic SQL 7-4


E

EMPTY_BLOB 8-24

EMPTY_CLOB 8-24

END 1-7

environments 1-13

EXCEPTION 1-7, 2-21

EXECUTE 4-3, 7-11

external large object 8-8

F

fetch 7-5

FILE_LOCATOR 8-16

file_type 7-27

formal parameter 2-6

forward declaration 6-8

function 3-3, 8-12

H

host variable 2-14

I

IDEPTREE 11-8

IMMEDIATE 7-11

INSTEAD OF 9-22

INSTEAD OF 9-7

internal 8-6

invoke a procedure 2-9

IS_OPEN 7-26

L

LOB 8-3, 8-5, 8-32

LOB locator 8-5

local dependencies 11-5

locator 8-12

LONG 8-4

LONG-to-LOB 8-17

M

migration 8-17

modularization 1-6

modules 1-6

mutating table 10-8

N

NEW 9-19


O

object privilege 4-3

OCI 8-10

OLD 9-19

one-time-only procedure 6-10

OPEN_CONNECTION 7-31

Oracle Internet Platform 1-4

overload 6-3

OLD and NEW qualifiers 9-19

P

package 4-16, 5-3, 7-22, 8-9, 8-19

package body 5-11

package specification 5-8

parameter mode 2-8, 3-8

Parsing 7-5

persistent state 6-14

PL/SQL block 2-4

PL/SQL construct 1-5

PROCEDURE 2-5, 2-25

procedures and functions within the 7-23

purity level 6-11

R

recompile a PL/SQL object 11-22

remote dependencies 11-12

row trigger 9-18

READ 8-12

REPLACE 2-4

REQUEST 7-29

REQUEST_PIECES 7-29

RETURN 3-4

row trigger 9-9

S

schedule batch job 7-13

security mechanism 8-9

SESSION_MAX_OPEN_FILE 8-13

SHOW ERROR 4-11


signature 11-13

SQL*Plus 1-4

statement trigger 9-9, 9-14

SUBMIT 7-15

submit PL/SQL program 7-13

subprogram 1-6

system event 10-3

system privileges 4-3

T

temporary 8-32

time stamp 11-13, 11-20

TO_BLOB 8-18

TO_CLOB 8-18

trigger action 9-10

trigger name 9-13

trigger timing 9-6

trigger type 9-6

triggering event 9-8

U

user-defined PL/SQL function 3-13

USER_DEPENDENCIES 11-7

USER_ERRORS 4-11

USER_JOBS 7-19

USER_OBJECTS 4-7

USER_SOURCE 4-9

USER_TRIGGER 10-28

UTL_FILE 7-21

UTL_FILE_DIR 7-22

UTL_HTTP 7-29

UTL_TCP 7-31

plsql 9i vol2

Technology

declaring plsql variables

plsql blocks

benefits of plsql

nonplsql variables

plsql environment

plsql records

plsql block structure

types of variables